mirrors/linux-stable
1
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git synced 2024-08-23 09:19:51 +00:00
Code Issues Releases Wiki Activity

perf/core improvements and fixes:

Browse source 
perf record:
 
   Alexey Budankov:
 
   - Implement --mmap-flush=<number> option, to control a threshold for draining
     the mmap ring buffers and consequently the size of the write calls to the
     output, be it perf.data, pipe mode or soon a compressor that with bigger
     buffers will do a better job before dumping compressed data into a new
     perf.data content mode, which is in the final steps of reviewing and testing.
 
 perf trace:
 
   Arnaldo Carvalho de Melo:
 
   - Add 'string' event alias to select syscalls with string args, i.e. for testing
     the BPF program used to copy those strings, allow for:
 
     # perf trace -e string
 
     To select all the syscalls that have things like pathnames.
 
   - Use a PERCPU_ARRAY BPF map to copy more string bytes than what is possible using
     the BPF stack, just like pioneered by the sysdig tool.
 
 Feature detection:
 
   Alexey Budankov:
 
   - Implement libzstd feature check, which is a library that provides a uniform
     API to various compression formats, will be used in 'perf record', see note
     about --mmap-flush feature.
 
 perf stat:
 
   Andi Kleen:
 
   - Implement a tool specific 'duration_time' event to allow showing the "time
     elapsed" line in the default 'perf stat' output as one of the events that
     can be asked for when using --field-separator and other script consumable
     outputs.
 
 Intel vendor events (JSON files):
 
   Andi Kleen:
 
   - Update metrics from TMAM 3.5.
 
   - Update events:
 
     Bonnell to V4
     Broadwell-DE to v7
     Broadwell to v23
     BroadwellX to v14
     GoldmontPlus to v1.01
     Goldmont to v13
     Haswell to v28
     HaswellX to v20
     IvyBridge to v21
     IvyTown to v20
     JakeTown to v20
     KnightsLanding to v9
     SandyBridge to v16
     Silvermont to v14
     Skylake to v42
     SkylakeX to v1.12
 
 IBM S/390 vendor events (JSON):
 
   Thomas Richter:
 
   - Fix s390 counter long description for L1D_RO_EXCL_WRITES.
 
 tools lib traceevent:
 
   Steven Rostedt (Red Hat):
 
   - Add more debugging to see various internal ring buffer entries.
 
   Steven Rostedt (VMWare):
 
   - Handle trace_printk() "%px".
   - Add mono clocks to be parsed in seconds.
   - Removed unneeded !! and return parenthesis.
 
   Tzvetomir Stoyanov :
 
   - Implement a new API, tep_list_events_copy().
   - Implement new traceevent APIs for accessing struct tep_handler fields.
   - Remove tep filter trivial APIs, not used anymore.
   - Remove call to exit() from tep_filter_add_filter_str(), library routines shouldn't
     kill tools using it.
   - Make traceevent APIs more consistent.
 
 Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYIAB0WIQR2GiIUctdOfX2qHhGyPKLppCJ+JwUCXKNwXgAKCRCyPKLppCJ+
 JypjAPsGYcZF1YFO5053kU6yo0fkVJB/3gyGQVjlAGqXcjsRnAD/cTnhni0ocQDW
 hu5nRBYCnw0l4r6yfg6Y6+jXlvCZyg8=
 =wiIr
 -----END PGP SIGNATURE-----

Merge tag 'perf-core-for-mingo-5.2-20190402' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core

Pull perf/core improvements and fixes from Arnaldo:

perf record:

  Alexey Budankov:

  - Implement --mmap-flush=<number> option, to control a threshold for draining
    the mmap ring buffers and consequently the size of the write calls to the
    output, be it perf.data, pipe mode or soon a compressor that with bigger
    buffers will do a better job before dumping compressed data into a new
    perf.data content mode, which is in the final steps of reviewing and testing.

perf trace:

  Arnaldo Carvalho de Melo:

  - Add 'string' event alias to select syscalls with string args, i.e. for testing
    the BPF program used to copy those strings, allow for:

    # perf trace -e string

    To select all the syscalls that have things like pathnames.

  - Use a PERCPU_ARRAY BPF map to copy more string bytes than what is possible using
    the BPF stack, just like pioneered by the sysdig tool.

Feature detection:

  Alexey Budankov:

  - Implement libzstd feature check, which is a library that provides a uniform
    API to various compression formats, will be used in 'perf record', see note
    about --mmap-flush feature.

perf stat:

  Andi Kleen:

  - Implement a tool specific 'duration_time' event to allow showing the "time
    elapsed" line in the default 'perf stat' output as one of the events that
    can be asked for when using --field-separator and other script consumable
    outputs.

Intel vendor events (JSON files):

  Andi Kleen:

  - Update metrics from TMAM 3.5.

  - Update events:

    Bonnell to V4
    Broadwell-DE to v7
    Broadwell to v23
    BroadwellX to v14
    GoldmontPlus to v1.01
    Goldmont to v13
    Haswell to v28
    HaswellX to v20
    IvyBridge to v21
    IvyTown to v20
    JakeTown to v20
    KnightsLanding to v9
    SandyBridge to v16
    Silvermont to v14
    Skylake to v42
    SkylakeX to v1.12

IBM S/390 vendor events (JSON):

  Thomas Richter:

  - Fix s390 counter long description for L1D_RO_EXCL_WRITES.

tools lib traceevent:

  Steven Rostedt (Red Hat):

  - Add more debugging to see various internal ring buffer entries.

  Steven Rostedt (VMWare):

  - Handle trace_printk() "%px".
  - Add mono clocks to be parsed in seconds.
  - Removed unneeded !! and return parenthesis.

  Tzvetomir Stoyanov :

  - Implement a new API, tep_list_events_copy().
  - Implement new traceevent APIs for accessing struct tep_handler fields.
  - Remove tep filter trivial APIs, not used anymore.
  - Remove call to exit() from tep_filter_add_filter_str(), library routines shouldn't
    kill tools using it.
  - Make traceevent APIs more consistent.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This commit is contained in:
Thomas Gleixner 2019-04-05 13:28:15 +02:00
commit cabf5ebbab
120 changed files with 12572 additions and 8275 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
tools/build/Makefile.feature
View file

@ -67,6 +67,7 @@ FEATURE_TESTS_BASIC := \
sdt \
setns \
libaio \
libzstd \
disassembler-four-args
# FEATURE_TESTS_BASIC + FEATURE_TESTS_EXTRA is the complete list
@ -120,6 +121,7 @@ FEATURE_DISPLAY ?= \
get_cpuid \
bpf \
libaio \
libzstd \
disassembler-four-args
# Set FEATURE_CHECK_(C|LD)FLAGS-all for all FEATURE_TESTS features.

6
tools/build/feature/Makefile
View file

@ -62,7 +62,8 @@ FILES= \
test-clang.bin \
test-llvm.bin \
test-llvm-version.bin \
test-libaio.bin
test-libaio.bin \
test-libzstd.bin
FILES := $(addprefix $(OUTPUT),$(FILES))
@ -301,6 +302,9 @@ $(OUTPUT)test-clang.bin:
$(OUTPUT)test-libaio.bin:
$(BUILD) -lrt
$(OUTPUT)test-libzstd.bin:
$(BUILD) -lzstd
###############################
clean:

5
tools/build/feature/test-all.c
View file

@ -182,6 +182,10 @@
# include "test-disassembler-four-args.c"
#undef main
#define main main_test_zstd
# include "test-libzstd.c"
#undef main
int main(int argc, char *argv[])
{
main_test_libpython();
@ -224,6 +228,7 @@ int main(int argc, char *argv[])
main_test_libaio();
main_test_reallocarray();
main_test_disassembler_four_args();
main_test_libzstd();
return 0;
}

12
tools/build/feature/test-libzstd.c Normal file
View file

@ -0,0 +1,12 @@
// SPDX-License-Identifier: GPL-2.0
#include <zstd.h>
int main(void)
{
ZSTD_CStream *cstream;
cstream = ZSTD_createCStream();
ZSTD_freeCStream(cstream);
return 0;
}

278
tools/lib/traceevent/event-parse-api.c
View file

@ -8,6 +8,22 @@
#include "event-parse-local.h"
#include "event-utils.h"
/**
* tep_get_event - returns the event with the given index
* @tep: a handle to the tep_handle
* @index: index of the requested event, in the range 0 .. nr_events
*
* This returns pointer to the element of the events array with the given index
* If @tep is NULL, or @index is not in the range 0 .. nr_events, NULL is returned.
*/
struct tep_event *tep_get_event(struct tep_handle *tep, int index)
{
if (tep && tep->events && index < tep->nr_events)
return tep->events[index];
return NULL;
}
/**
* tep_get_first_event - returns the first event in the events array
* @tep: a handle to the tep_handle
@ -17,10 +33,7 @@
*/
struct tep_event *tep_get_first_event(struct tep_handle *tep)
{
if (tep && tep->events)
return tep->events[0];
return NULL;
return tep_get_event(tep, 0);
}
/**
@ -32,7 +45,7 @@ struct tep_event *tep_get_first_event(struct tep_handle *tep)
*/
int tep_get_events_count(struct tep_handle *tep)
{
if(tep)
if (tep)
return tep->nr_events;
return 0;
}
@ -43,19 +56,47 @@ int tep_get_events_count(struct tep_handle *tep)
* @flag: flag, or combination of flags to be set
* can be any combination from enum tep_flag
*
* This sets a flag or mbination of flags from enum tep_flag
*/
* This sets a flag or combination of flags from enum tep_flag
*/
void tep_set_flag(struct tep_handle *tep, int flag)
{
if(tep)
if (tep)
tep->flags |= flag;
}
unsigned short tep_data2host2(struct tep_handle *pevent, unsigned short data)
/**
* tep_clear_flag - clear event parser flag
* @tep: a handle to the tep_handle
* @flag: flag to be cleared
*
* This clears a tep flag
*/
void tep_clear_flag(struct tep_handle *tep, enum tep_flag flag)
{
if (tep)
tep->flags &= ~flag;
}
/**
* tep_test_flag - check the state of event parser flag
* @tep: a handle to the tep_handle
* @flag: flag to be checked
*
* This returns the state of the requested tep flag.
* Returns: true if the flag is set, false otherwise.
*/
bool tep_test_flag(struct tep_handle *tep, enum tep_flag flag)
{
if (tep)
return tep->flags & flag;
return false;
}
unsigned short tep_data2host2(struct tep_handle *tep, unsigned short data)
{
unsigned short swap;
if (!pevent || pevent->host_bigendian == pevent->file_bigendian)
if (!tep || tep->host_bigendian == tep->file_bigendian)
return data;
swap = ((data & 0xffULL) << 8) |
@ -64,11 +105,11 @@ unsigned short tep_data2host2(struct tep_handle *pevent, unsigned short data)
return swap;
}
unsigned int tep_data2host4(struct tep_handle *pevent, unsigned int data)
unsigned int tep_data2host4(struct tep_handle *tep, unsigned int data)
{
unsigned int swap;
if (!pevent || pevent->host_bigendian == pevent->file_bigendian)
if (!tep || tep->host_bigendian == tep->file_bigendian)
return data;
swap = ((data & 0xffULL) << 24) |
@ -80,11 +121,11 @@ unsigned int tep_data2host4(struct tep_handle *pevent, unsigned int data)
}
unsigned long long
tep_data2host8(struct tep_handle *pevent, unsigned long long data)
tep_data2host8(struct tep_handle *tep, unsigned long long data)
{
unsigned long long swap;
if (!pevent || pevent->host_bigendian == pevent->file_bigendian)
if (!tep || tep->host_bigendian == tep->file_bigendian)
return data;
swap = ((data & 0xffULL) << 56) |
@ -101,175 +142,232 @@ tep_data2host8(struct tep_handle *pevent, unsigned long long data)
/**
* tep_get_header_page_size - get size of the header page
* @pevent: a handle to the tep_handle
* @tep: a handle to the tep_handle
*
* This returns size of the header page
* If @pevent is NULL, 0 is returned.
* If @tep is NULL, 0 is returned.
*/
int tep_get_header_page_size(struct tep_handle *pevent)
int tep_get_header_page_size(struct tep_handle *tep)
{
if(pevent)
return pevent->header_page_size_size;
if (tep)
return tep->header_page_size_size;
return 0;
}
/**
* tep_get_header_timestamp_size - get size of the timestamp in the header page
* @tep: a handle to the tep_handle
*
* This returns size of the timestamp in the header page
* If @tep is NULL, 0 is returned.
*/
int tep_get_header_timestamp_size(struct tep_handle *tep)
{
if (tep)
return tep->header_page_ts_size;
return 0;
}
/**
* tep_get_cpus - get the number of CPUs
* @pevent: a handle to the tep_handle
* @tep: a handle to the tep_handle
*
* This returns the number of CPUs
* If @pevent is NULL, 0 is returned.
* If @tep is NULL, 0 is returned.
*/
int tep_get_cpus(struct tep_handle *pevent)
int tep_get_cpus(struct tep_handle *tep)
{
if(pevent)
return pevent->cpus;
if (tep)
return tep->cpus;
return 0;
}
/**
* tep_set_cpus - set the number of CPUs
* @pevent: a handle to the tep_handle
* @tep: a handle to the tep_handle
*
* This sets the number of CPUs
*/
void tep_set_cpus(struct tep_handle *pevent, int cpus)
void tep_set_cpus(struct tep_handle *tep, int cpus)
{
if(pevent)
pevent->cpus = cpus;
if (tep)
tep->cpus = cpus;
}
/**
* tep_get_long_size - get the size of a long integer on the current machine
* @pevent: a handle to the tep_handle
* tep_get_long_size - get the size of a long integer on the traced machine
* @tep: a handle to the tep_handle
*
* This returns the size of a long integer on the current machine
* If @pevent is NULL, 0 is returned.
* This returns the size of a long integer on the traced machine
* If @tep is NULL, 0 is returned.
*/
int tep_get_long_size(struct tep_handle *pevent)
int tep_get_long_size(struct tep_handle *tep)
{
if(pevent)
return pevent->long_size;
if (tep)
return tep->long_size;
return 0;
}
/**
* tep_set_long_size - set the size of a long integer on the current machine
* @pevent: a handle to the tep_handle
* tep_set_long_size - set the size of a long integer on the traced machine
* @tep: a handle to the tep_handle
* @size: size, in bytes, of a long integer
*
* This sets the size of a long integer on the current machine
* This sets the size of a long integer on the traced machine
*/
void tep_set_long_size(struct tep_handle *pevent, int long_size)
void tep_set_long_size(struct tep_handle *tep, int long_size)
{
if(pevent)
pevent->long_size = long_size;
if (tep)
tep->long_size = long_size;
}
/**
* tep_get_page_size - get the size of a memory page on the current machine
* @pevent: a handle to the tep_handle
* tep_get_page_size - get the size of a memory page on the traced machine
* @tep: a handle to the tep_handle
*
* This returns the size of a memory page on the current machine
* If @pevent is NULL, 0 is returned.
* This returns the size of a memory page on the traced machine
* If @tep is NULL, 0 is returned.
*/
int tep_get_page_size(struct tep_handle *pevent)
int tep_get_page_size(struct tep_handle *tep)
{
if(pevent)
return pevent->page_size;
if (tep)
return tep->page_size;
return 0;
}
/**
* tep_set_page_size - set the size of a memory page on the current machine
* @pevent: a handle to the tep_handle
* tep_set_page_size - set the size of a memory page on the traced machine
* @tep: a handle to the tep_handle
* @_page_size: size of a memory page, in bytes
*
* This sets the size of a memory page on the current machine
* This sets the size of a memory page on the traced machine
*/
void tep_set_page_size(struct tep_handle *pevent, int _page_size)
void tep_set_page_size(struct tep_handle *tep, int _page_size)
{
if(pevent)
pevent->page_size = _page_size;
if (tep)
tep->page_size = _page_size;
}
/**
* tep_file_bigendian - get if the file is in big endian order
* @pevent: a handle to the tep_handle
* tep_is_file_bigendian - return the endian of the file
* @tep: a handle to the tep_handle
*
* This returns if the file is in big endian order
* If @pevent is NULL, 0 is returned.
* This returns true if the file is in big endian order
* If @tep is NULL, false is returned.
*/
int tep_file_bigendian(struct tep_handle *pevent)
bool tep_is_file_bigendian(struct tep_handle *tep)
{
if(pevent)
return pevent->file_bigendian;
return 0;
if (tep)
return (tep->file_bigendian == TEP_BIG_ENDIAN);
return false;
}
/**
* tep_set_file_bigendian - set if the file is in big endian order
* @pevent: a handle to the tep_handle
* @tep: a handle to the tep_handle
* @endian: non zero, if the file is in big endian order
*
* This sets if the file is in big endian order
*/
void tep_set_file_bigendian(struct tep_handle *pevent, enum tep_endian endian)
void tep_set_file_bigendian(struct tep_handle *tep, enum tep_endian endian)
{
if(pevent)
pevent->file_bigendian = endian;
if (tep)
tep->file_bigendian = endian;
}
/**
* tep_is_host_bigendian - get if the order of the current host is big endian
* @pevent: a handle to the tep_handle
* tep_is_local_bigendian - return the endian of the saved local machine
* @tep: a handle to the tep_handle
*
* This gets if the order of the current host is big endian
* If @pevent is NULL, 0 is returned.
* This returns true if the saved local machine in @tep is big endian.
* If @tep is NULL, false is returned.
*/
int tep_is_host_bigendian(struct tep_handle *pevent)
bool tep_is_local_bigendian(struct tep_handle *tep)
{
if(pevent)
return pevent->host_bigendian;
if (tep)
return (tep->host_bigendian == TEP_BIG_ENDIAN);
return 0;
}
/**
* tep_set_host_bigendian - set the order of the local host
* @pevent: a handle to the tep_handle
* tep_set_local_bigendian - set the stored local machine endian order
* @tep: a handle to the tep_handle
* @endian: non zero, if the local host has big endian order
*
* This sets the order of the local host
* This sets the endian order for the local machine.
*/
void tep_set_host_bigendian(struct tep_handle *pevent, enum tep_endian endian)
void tep_set_local_bigendian(struct tep_handle *tep, enum tep_endian endian)
{
if(pevent)
pevent->host_bigendian = endian;
if (tep)
tep->host_bigendian = endian;
}
/**
* tep_is_latency_format - get if the latency output format is configured
* @pevent: a handle to the tep_handle
* @tep: a handle to the tep_handle
*
* This gets if the latency output format is configured
* If @pevent is NULL, 0 is returned.
* This returns true if the latency output format is configured
* If @tep is NULL, false is returned.
*/
int tep_is_latency_format(struct tep_handle *pevent)
bool tep_is_latency_format(struct tep_handle *tep)
{
if(pevent)
return pevent->latency_format;
return 0;
if (tep)
return (tep->latency_format);
return false;
}
/**
* tep_set_latency_format - set the latency output format
* @pevent: a handle to the tep_handle
* @tep: a handle to the tep_handle
* @lat: non zero for latency output format
*
* This sets the latency output format
*/
void tep_set_latency_format(struct tep_handle *pevent, int lat)
void tep_set_latency_format(struct tep_handle *tep, int lat)
{
if(pevent)
pevent->latency_format = lat;
if (tep)
tep->latency_format = lat;
}
/**
* tep_is_old_format - get if an old kernel is used
* @tep: a handle to the tep_handle
*
* This returns true, if an old kernel is used to generate the tracing events or
* false if a new kernel is used. Old kernels did not have header page info.
* If @tep is NULL, false is returned.
*/
bool tep_is_old_format(struct tep_handle *tep)
{
if (tep)
return tep->old_format;
return false;
}
/**
* tep_set_print_raw - set a flag to force print in raw format
* @tep: a handle to the tep_handle
* @print_raw: the new value of the print_raw flag
*
* This sets a flag to force print in raw format
*/
void tep_set_print_raw(struct tep_handle *tep, int print_raw)
{
if (tep)
tep->print_raw = print_raw;
}
/**
* tep_set_test_filters - set a flag to test a filter string
* @tep: a handle to the tep_handle
* @test_filters: the new value of the test_filters flag
*
* This sets a flag to test a filter string. If this flag is set, when
* tep_filter_add_filter_str() API as called,it will print the filter string
* instead of adding it.
*/
void tep_set_test_filters(struct tep_handle *tep, int test_filters)
{
if (tep)
tep->test_filters = test_filters;
}

6
tools/lib/traceevent/event-parse-local.h
View file

@ -92,8 +92,8 @@ struct tep_handle {
void tep_free_event(struct tep_event *event);
void tep_free_format_field(struct tep_format_field *field);
unsigned short tep_data2host2(struct tep_handle *pevent, unsigned short data);
unsigned int tep_data2host4(struct tep_handle *pevent, unsigned int data);
unsigned long long tep_data2host8(struct tep_handle *pevent, unsigned long long data);
unsigned short tep_data2host2(struct tep_handle *tep, unsigned short data);
unsigned int tep_data2host4(struct tep_handle *tep, unsigned int data);
unsigned long long tep_data2host8(struct tep_handle *tep, unsigned long long data);
#endif /* _PARSE_EVENTS_INT_H */

917
tools/lib/traceevent/event-parse.c
View file

File diff suppressed because it is too large Load diff

152
tools/lib/traceevent/event-parse.h
View file

@ -64,8 +64,8 @@ typedef int (*tep_event_handler_func)(struct trace_seq *s,
struct tep_event *event,
void *context);
typedef int (*tep_plugin_load_func)(struct tep_handle *pevent);
typedef int (*tep_plugin_unload_func)(struct tep_handle *pevent);
typedef int (*tep_plugin_load_func)(struct tep_handle *tep);
typedef int (*tep_plugin_unload_func)(struct tep_handle *tep);
struct tep_plugin_option {
struct tep_plugin_option *next;
@ -85,12 +85,12 @@ struct tep_plugin_option {
* TEP_PLUGIN_LOADER: (required)
* The function name to initialized the plugin.
*
* int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
* int TEP_PLUGIN_LOADER(struct tep_handle *tep)
*
* TEP_PLUGIN_UNLOADER: (optional)
* The function called just before unloading
*
* int TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
* int TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
*
* TEP_PLUGIN_OPTIONS: (optional)
* Plugin options that can be set before loading
@ -278,7 +278,7 @@ struct tep_print_fmt {
};
struct tep_event {
struct tep_handle *pevent;
struct tep_handle *tep;
char *name;
int id;
int flags;
@ -393,9 +393,9 @@ struct tep_plugin_list;
#define INVALID_PLUGIN_LIST_OPTION ((char **)((unsigned long)-1))
struct tep_plugin_list *tep_load_plugins(struct tep_handle *pevent);
struct tep_plugin_list *tep_load_plugins(struct tep_handle *tep);
void tep_unload_plugins(struct tep_plugin_list *plugin_list,
struct tep_handle *pevent);
struct tep_handle *tep);
char **tep_plugin_list_options(void);
void tep_plugin_free_options_list(char **list);
int tep_plugin_add_options(const char *name,
@ -409,8 +409,10 @@ void tep_print_plugins(struct trace_seq *s,
typedef char *(tep_func_resolver_t)(void *priv,
unsigned long long *addrp, char **modp);
void tep_set_flag(struct tep_handle *tep, int flag);
void tep_clear_flag(struct tep_handle *tep, enum tep_flag flag);
bool tep_test_flag(struct tep_handle *tep, enum tep_flag flags);
static inline int tep_host_bigendian(void)
static inline int tep_is_bigendian(void)
{
unsigned char str[] = { 0x1, 0x2, 0x3, 0x4 };
unsigned int val;
@ -428,37 +430,37 @@ enum trace_flag_type {
TRACE_FLAG_SOFTIRQ = 0x10,
};
int tep_set_function_resolver(struct tep_handle *pevent,
int tep_set_function_resolver(struct tep_handle *tep,
tep_func_resolver_t *func, void *priv);
void tep_reset_function_resolver(struct tep_handle *pevent);
int tep_register_comm(struct tep_handle *pevent, const char *comm, int pid);
int tep_override_comm(struct tep_handle *pevent, const char *comm, int pid);
int tep_register_trace_clock(struct tep_handle *pevent, const char *trace_clock);
int tep_register_function(struct tep_handle *pevent, char *name,
void tep_reset_function_resolver(struct tep_handle *tep);
int tep_register_comm(struct tep_handle *tep, const char *comm, int pid);
int tep_override_comm(struct tep_handle *tep, const char *comm, int pid);
int tep_register_trace_clock(struct tep_handle *tep, const char *trace_clock);
int tep_register_function(struct tep_handle *tep, char *name,
unsigned long long addr, char *mod);
int tep_register_print_string(struct tep_handle *pevent, const char *fmt,
int tep_register_print_string(struct tep_handle *tep, const char *fmt,
unsigned long long addr);
int tep_pid_is_registered(struct tep_handle *pevent, int pid);
bool tep_is_pid_registered(struct tep_handle *tep, int pid);
void tep_print_event_task(struct tep_handle *pevent, struct trace_seq *s,
void tep_print_event_task(struct tep_handle *tep, struct trace_seq *s,
struct tep_event *event,
struct tep_record *record);
void tep_print_event_time(struct tep_handle *pevent, struct trace_seq *s,
void tep_print_event_time(struct tep_handle *tep, struct trace_seq *s,
struct tep_event *event,
struct tep_record *record,
bool use_trace_clock);
void tep_print_event_data(struct tep_handle *pevent, struct trace_seq *s,
void tep_print_event_data(struct tep_handle *tep, struct trace_seq *s,
struct tep_event *event,
struct tep_record *record);
void tep_print_event(struct tep_handle *pevent, struct trace_seq *s,
void tep_print_event(struct tep_handle *tep, struct trace_seq *s,
struct tep_record *record, bool use_trace_clock);
int tep_parse_header_page(struct tep_handle *pevent, char *buf, unsigned long size,
int tep_parse_header_page(struct tep_handle *tep, char *buf, unsigned long size,
int long_size);
enum tep_errno tep_parse_event(struct tep_handle *pevent, const char *buf,
enum tep_errno tep_parse_event(struct tep_handle *tep, const char *buf,
unsigned long size, const char *sys);
enum tep_errno tep_parse_format(struct tep_handle *pevent,
enum tep_errno tep_parse_format(struct tep_handle *tep,
struct tep_event **eventp,
const char *buf,
unsigned long size, const char *sys);
@ -490,50 +492,50 @@ enum tep_reg_handler {
TEP_REGISTER_SUCCESS_OVERWRITE,
};
int tep_register_event_handler(struct tep_handle *pevent, int id,
int tep_register_event_handler(struct tep_handle *tep, int id,
const char *sys_name, const char *event_name,
tep_event_handler_func func, void *context);
int tep_unregister_event_handler(struct tep_handle *pevent, int id,
int tep_unregister_event_handler(struct tep_handle *tep, int id,
const char *sys_name, const char *event_name,
tep_event_handler_func func, void *context);
int tep_register_print_function(struct tep_handle *pevent,
int tep_register_print_function(struct tep_handle *tep,
tep_func_handler func,
enum tep_func_arg_type ret_type,
char *name, ...);
int tep_unregister_print_function(struct tep_handle *pevent,
int tep_unregister_print_function(struct tep_handle *tep,
tep_func_handler func, char *name);
struct tep_format_field *tep_find_common_field(struct tep_event *event, const char *name);
struct tep_format_field *tep_find_field(struct tep_event *event, const char *name);
struct tep_format_field *tep_find_any_field(struct tep_event *event, const char *name);
const char *tep_find_function(struct tep_handle *pevent, unsigned long long addr);
const char *tep_find_function(struct tep_handle *tep, unsigned long long addr);
unsigned long long
tep_find_function_address(struct tep_handle *pevent, unsigned long long addr);
unsigned long long tep_read_number(struct tep_handle *pevent, const void *ptr, int size);
tep_find_function_address(struct tep_handle *tep, unsigned long long addr);
unsigned long long tep_read_number(struct tep_handle *tep, const void *ptr, int size);
int tep_read_number_field(struct tep_format_field *field, const void *data,
unsigned long long *value);
struct tep_event *tep_get_first_event(struct tep_handle *tep);
int tep_get_events_count(struct tep_handle *tep);
struct tep_event *tep_find_event(struct tep_handle *pevent, int id);
struct tep_event *tep_find_event(struct tep_handle *tep, int id);
struct tep_event *
tep_find_event_by_name(struct tep_handle *pevent, const char *sys, const char *name);
tep_find_event_by_name(struct tep_handle *tep, const char *sys, const char *name);
struct tep_event *
tep_find_event_by_record(struct tep_handle *pevent, struct tep_record *record);
tep_find_event_by_record(struct tep_handle *tep, struct tep_record *record);
void tep_data_lat_fmt(struct tep_handle *pevent,
struct trace_seq *s, struct tep_record *record);
int tep_data_type(struct tep_handle *pevent, struct tep_record *rec);
int tep_data_pid(struct tep_handle *pevent, struct tep_record *rec);
int tep_data_preempt_count(struct tep_handle *pevent, struct tep_record *rec);
int tep_data_flags(struct tep_handle *pevent, struct tep_record *rec);
const char *tep_data_comm_from_pid(struct tep_handle *pevent, int pid);
void tep_data_latency_format(struct tep_handle *tep,
struct trace_seq *s, struct tep_record *record);
int tep_data_type(struct tep_handle *tep, struct tep_record *rec);
int tep_data_pid(struct tep_handle *tep, struct tep_record *rec);
int tep_data_preempt_count(struct tep_handle *tep, struct tep_record *rec);
int tep_data_flags(struct tep_handle *tep, struct tep_record *rec);
const char *tep_data_comm_from_pid(struct tep_handle *tep, int pid);
struct tep_cmdline;
struct tep_cmdline *tep_data_pid_from_comm(struct tep_handle *pevent, const char *comm,
struct tep_cmdline *tep_data_pid_from_comm(struct tep_handle *tep, const char *comm,
struct tep_cmdline *next);
int tep_cmdline_pid(struct tep_handle *pevent, struct tep_cmdline *cmdline);
int tep_cmdline_pid(struct tep_handle *tep, struct tep_cmdline *cmdline);
void tep_print_field(struct trace_seq *s, void *data,
struct tep_format_field *field);
@ -541,10 +543,12 @@ void tep_print_fields(struct trace_seq *s, void *data,
int size __maybe_unused, struct tep_event *event);
void tep_event_info(struct trace_seq *s, struct tep_event *event,
struct tep_record *record);
int tep_strerror(struct tep_handle *pevent, enum tep_errno errnum,
int tep_strerror(struct tep_handle *tep, enum tep_errno errnum,
char *buf, size_t buflen);
struct tep_event **tep_list_events(struct tep_handle *pevent, enum tep_event_sort_type);
struct tep_event **tep_list_events(struct tep_handle *tep, enum tep_event_sort_type);
struct tep_event **tep_list_events_copy(struct tep_handle *tep,
enum tep_event_sort_type);
struct tep_format_field **tep_event_common_fields(struct tep_event *event);
struct tep_format_field **tep_event_fields(struct tep_event *event);
@ -552,24 +556,28 @@ enum tep_endian {
TEP_LITTLE_ENDIAN = 0,
TEP_BIG_ENDIAN
};
int tep_get_cpus(struct tep_handle *pevent);
void tep_set_cpus(struct tep_handle *pevent, int cpus);
int tep_get_long_size(struct tep_handle *pevent);
void tep_set_long_size(struct tep_handle *pevent, int long_size);
int tep_get_page_size(struct tep_handle *pevent);
void tep_set_page_size(struct tep_handle *pevent, int _page_size);
int tep_file_bigendian(struct tep_handle *pevent);
void tep_set_file_bigendian(struct tep_handle *pevent, enum tep_endian endian);
int tep_is_host_bigendian(struct tep_handle *pevent);
void tep_set_host_bigendian(struct tep_handle *pevent, enum tep_endian endian);
int tep_is_latency_format(struct tep_handle *pevent);
void tep_set_latency_format(struct tep_handle *pevent, int lat);
int tep_get_header_page_size(struct tep_handle *pevent);
int tep_get_cpus(struct tep_handle *tep);
void tep_set_cpus(struct tep_handle *tep, int cpus);
int tep_get_long_size(struct tep_handle *tep);
void tep_set_long_size(struct tep_handle *tep, int long_size);
int tep_get_page_size(struct tep_handle *tep);
void tep_set_page_size(struct tep_handle *tep, int _page_size);
bool tep_is_file_bigendian(struct tep_handle *tep);
void tep_set_file_bigendian(struct tep_handle *tep, enum tep_endian endian);
bool tep_is_local_bigendian(struct tep_handle *tep);
void tep_set_local_bigendian(struct tep_handle *tep, enum tep_endian endian);
bool tep_is_latency_format(struct tep_handle *tep);
void tep_set_latency_format(struct tep_handle *tep, int lat);
int tep_get_header_page_size(struct tep_handle *tep);
int tep_get_header_timestamp_size(struct tep_handle *tep);
bool tep_is_old_format(struct tep_handle *tep);
void tep_set_print_raw(struct tep_handle *tep, int print_raw);
void tep_set_test_filters(struct tep_handle *tep, int test_filters);
struct tep_handle *tep_alloc(void);
void tep_free(struct tep_handle *pevent);
void tep_ref(struct tep_handle *pevent);
void tep_unref(struct tep_handle *pevent);
void tep_free(struct tep_handle *tep);
void tep_ref(struct tep_handle *tep);
void tep_unref(struct tep_handle *tep);
int tep_get_ref(struct tep_handle *tep);
/* access to the internal parser */
@ -581,8 +589,8 @@ const char *tep_get_input_buf(void);
unsigned long long tep_get_input_buf_ptr(void);
/* for debugging */
void tep_print_funcs(struct tep_handle *pevent);
void tep_print_printk(struct tep_handle *pevent);
void tep_print_funcs(struct tep_handle *tep);
void tep_print_printk(struct tep_handle *tep);
/* ----------------------- filtering ----------------------- */
@ -709,13 +717,13 @@ struct tep_filter_type {
#define TEP_FILTER_ERROR_BUFSZ 1024
struct tep_event_filter {
struct tep_handle *pevent;
struct tep_handle *tep;
int filters;
struct tep_filter_type *event_filters;
char error_buffer[TEP_FILTER_ERROR_BUFSZ];
};
struct tep_event_filter *tep_filter_alloc(struct tep_handle *pevent);
struct tep_event_filter *tep_filter_alloc(struct tep_handle *tep);
/* for backward compatibility */
#define FILTER_NONE TEP_ERRNO__NO_FILTER
@ -723,12 +731,6 @@ struct tep_event_filter *tep_filter_alloc(struct tep_handle *pevent);
#define FILTER_MISS TEP_ERRNO__FILTER_MISS
#define FILTER_MATCH TEP_ERRNO__FILTER_MATCH
enum tep_filter_trivial_type {
TEP_FILTER_TRIVIAL_FALSE,
TEP_FILTER_TRIVIAL_TRUE,
TEP_FILTER_TRIVIAL_BOTH,
};
enum tep_errno tep_filter_add_filter_str(struct tep_event_filter *filter,
const char *filter_str);
@ -743,9 +745,6 @@ int tep_event_filtered(struct tep_event_filter *filter,
void tep_filter_reset(struct tep_event_filter *filter);
int tep_filter_clear_trivial(struct tep_event_filter *filter,
enum tep_filter_trivial_type type);
void tep_filter_free(struct tep_event_filter *filter);
char *tep_filter_make_string(struct tep_event_filter *filter, int event_id);
@ -753,15 +752,8 @@ char *tep_filter_make_string(struct tep_event_filter *filter, int event_id);
int tep_filter_remove_event(struct tep_event_filter *filter,
int event_id);
int tep_filter_event_has_trivial(struct tep_event_filter *filter,
int event_id,
enum tep_filter_trivial_type type);
int tep_filter_copy(struct tep_event_filter *dest, struct tep_event_filter *source);
int tep_update_trivial(struct tep_event_filter *dest, struct tep_event_filter *source,
enum tep_filter_trivial_type type);
int tep_filter_compare(struct tep_event_filter *filter1, struct tep_event_filter *filter2);
#endif /* _PARSE_EVENTS_H */

32
tools/lib/traceevent/event-plugin.c
View file

@ -269,7 +269,7 @@ void tep_print_plugins(struct trace_seq *s,
}
static void
load_plugin(struct tep_handle *pevent, const char *path,
load_plugin(struct tep_handle *tep, const char *path,
const char *file, void *data)
{
struct tep_plugin_list **plugin_list = data;
@ -316,7 +316,7 @@ load_plugin(struct tep_handle *pevent, const char *path,
*plugin_list = list;
pr_stat("registering plugin: %s", plugin);
func(pevent);
func(tep);
return;
out_free:
@ -324,9 +324,9 @@ load_plugin(struct tep_handle *pevent, const char *path,
}
static void
load_plugins_dir(struct tep_handle *pevent, const char *suffix,
load_plugins_dir(struct tep_handle *tep, const char *suffix,
const char *path,
void (*load_plugin)(struct tep_handle *pevent,
void (*load_plugin)(struct tep_handle *tep,
const char *path,
const char *name,
void *data),
@ -359,15 +359,15 @@ load_plugins_dir(struct tep_handle *pevent, const char *suffix,
if (strcmp(name + (strlen(name) - strlen(suffix)), suffix) != 0)
continue;
load_plugin(pevent, path, name, data);
load_plugin(tep, path, name, data);
}
closedir(dir);
}
static void
load_plugins(struct tep_handle *pevent, const char *suffix,
void (*load_plugin)(struct tep_handle *pevent,
load_plugins(struct tep_handle *tep, const char *suffix,
void (*load_plugin)(struct tep_handle *tep,
const char *path,
const char *name,
void *data),
@ -378,7 +378,7 @@ load_plugins(struct tep_handle *pevent, const char *suffix,
char *envdir;
int ret;
if (pevent->flags & TEP_DISABLE_PLUGINS)
if (tep->flags & TEP_DISABLE_PLUGINS)
return;
/*
@ -386,8 +386,8 @@ load_plugins(struct tep_handle *pevent, const char *suffix,
* check that first.
*/
#ifdef PLUGIN_DIR
if (!(pevent->flags & TEP_DISABLE_SYS_PLUGINS))
load_plugins_dir(pevent, suffix, PLUGIN_DIR,
if (!(tep->flags & TEP_DISABLE_SYS_PLUGINS))
load_plugins_dir(tep, suffix, PLUGIN_DIR,
load_plugin, data);
#endif
@ -397,7 +397,7 @@ load_plugins(struct tep_handle *pevent, const char *suffix,
*/
envdir = getenv("TRACEEVENT_PLUGIN_DIR");
if (envdir)
load_plugins_dir(pevent, suffix, envdir, load_plugin, data);
load_plugins_dir(tep, suffix, envdir, load_plugin, data);
/*
* Now let the home directory override the environment
@ -413,22 +413,22 @@ load_plugins(struct tep_handle *pevent, const char *suffix,
return;
}
load_plugins_dir(pevent, suffix, path, load_plugin, data);
load_plugins_dir(tep, suffix, path, load_plugin, data);
free(path);
}
struct tep_plugin_list*
tep_load_plugins(struct tep_handle *pevent)
tep_load_plugins(struct tep_handle *tep)
{
struct tep_plugin_list *list = NULL;
load_plugins(pevent, ".so", load_plugin, &list);
load_plugins(tep, ".so", load_plugin, &list);
return list;
}
void
tep_unload_plugins(struct tep_plugin_list *plugin_list, struct tep_handle *pevent)
tep_unload_plugins(struct tep_plugin_list *plugin_list, struct tep_handle *tep)
{
tep_plugin_unload_func func;
struct tep_plugin_list *list;
@ -438,7 +438,7 @@ tep_unload_plugins(struct tep_plugin_list *plugin_list, struct tep_handle *peven
plugin_list = list->next;
func = dlsym(list->handle, TEP_PLUGIN_UNLOADER_NAME);
if (func)
func(pevent);
func(tep);
dlclose(list->handle);
free(list->name);
free(list);

49
tools/lib/traceevent/kbuffer-parse.c
View file

@ -727,3 +727,52 @@ int kbuffer_start_of_data(struct kbuffer *kbuf)
{
return kbuf->start;
}
/**
* kbuffer_raw_get - get raw buffer info
* @kbuf: The kbuffer
* @subbuf: Start of mapped subbuffer
* @info: Info descriptor to fill in
*
* For debugging. This can return internals of the ring buffer.
* Expects to have info->next set to what it will read.
* The type, length and timestamp delta will be filled in, and
* @info->next will be updated to the next element.
* The @subbuf is used to know if the info is passed the end of
* data and NULL will be returned if it is.
*/
struct kbuffer_raw_info *
kbuffer_raw_get(struct kbuffer *kbuf, void *subbuf, struct kbuffer_raw_info *info)
{
unsigned long long flags;
unsigned long long delta;
unsigned int type_len;
unsigned int size;
int start;
int length;
void *ptr = info->next;
if (!kbuf || !subbuf)
return NULL;
if (kbuf->flags & KBUFFER_FL_LONG_8)
start = 16;
else
start = 12;
flags = read_long(kbuf, subbuf + 8);
size = (unsigned int)flags & COMMIT_MASK;
if (ptr < subbuf || ptr >= subbuf + start + size)
return NULL;
type_len = translate_data(kbuf, ptr, &ptr, &delta, &length);
info->next = ptr + length;
info->type = type_len;
info->delta = delta;
info->length = length;
return info;
}

13
tools/lib/traceevent/kbuffer.h
View file

@ -65,4 +65,17 @@ int kbuffer_subbuffer_size(struct kbuffer *kbuf);
void kbuffer_set_old_format(struct kbuffer *kbuf);
int kbuffer_start_of_data(struct kbuffer *kbuf);
/* Debugging */
struct kbuffer_raw_info {
int type;
int length;
unsigned long long delta;
void *next;
};
/* Read raw data */
struct kbuffer_raw_info *kbuffer_raw_get(struct kbuffer *kbuf, void *subbuf,
struct kbuffer_raw_info *info);
#endif /* _K_BUFFER_H */

216
tools/lib/traceevent/parse-filter.c
View file

@ -154,7 +154,7 @@ add_filter_type(struct tep_event_filter *filter, int id)
filter_type = &filter->event_filters[i];
filter_type->event_id = id;
filter_type->event = tep_find_event(filter->pevent, id);
filter_type->event = tep_find_event(filter->tep, id);
filter_type->filter = NULL;
filter->filters++;
@ -164,9 +164,9 @@ add_filter_type(struct tep_event_filter *filter, int id)
/**
* tep_filter_alloc - create a new event filter
* @pevent: The pevent that this filter is associated with
* @tep: The tep that this filter is associated with
*/
struct tep_event_filter *tep_filter_alloc(struct tep_handle *pevent)
struct tep_event_filter *tep_filter_alloc(struct tep_handle *tep)
{
struct tep_event_filter *filter;
@ -175,8 +175,8 @@ struct tep_event_filter *tep_filter_alloc(struct tep_handle *pevent)
return NULL;
memset(filter, 0, sizeof(*filter));
filter->pevent = pevent;
tep_ref(pevent);
filter->tep = tep;
tep_ref(tep);
return filter;
}
@ -256,7 +256,7 @@ static int event_match(struct tep_event *event,
}
static enum tep_errno
find_event(struct tep_handle *pevent, struct event_list **events,
find_event(struct tep_handle *tep, struct event_list **events,
char *sys_name, char *event_name)
{
struct tep_event *event;
@ -299,8 +299,8 @@ find_event(struct tep_handle *pevent, struct event_list **events,
}
}
for (i = 0; i < pevent->nr_events; i++) {
event = pevent->events[i];
for (i = 0; i < tep->nr_events; i++) {
event = tep->events[i];
if (event_match(event, sys_name ? &sreg : NULL, &ereg)) {
match = 1;
if (add_event(events, event) < 0) {
@ -1257,7 +1257,7 @@ static void filter_init_error_buf(struct tep_event_filter *filter)
enum tep_errno tep_filter_add_filter_str(struct tep_event_filter *filter,
const char *filter_str)
{
struct tep_handle *pevent = filter->pevent;
struct tep_handle *tep = filter->tep;
struct event_list *event;
struct event_list *events = NULL;
const char *filter_start;
@ -1313,7 +1313,7 @@ enum tep_errno tep_filter_add_filter_str(struct tep_event_filter *filter,
}
/* Find this event */
ret = find_event(pevent, &events, strim(sys_name), strim(event_name));
ret = find_event(tep, &events, strim(sys_name), strim(event_name));
if (ret < 0) {
free_events(events);
free(this_event);
@ -1334,7 +1334,7 @@ enum tep_errno tep_filter_add_filter_str(struct tep_event_filter *filter,
if (ret < 0)
rtn = ret;
if (ret >= 0 && pevent->test_filters) {
if (ret >= 0 && tep->test_filters) {
char *test;
test = tep_filter_make_string(filter, event->event->id);
if (test) {
@ -1346,9 +1346,6 @@ enum tep_errno tep_filter_add_filter_str(struct tep_event_filter *filter,
free_events(events);
if (rtn >= 0 && pevent->test_filters)
exit(0);
return rtn;
}
@ -1380,7 +1377,7 @@ int tep_filter_strerror(struct tep_event_filter *filter, enum tep_errno err,
return 0;
}
return tep_strerror(filter->pevent, err, buf, buflen);
return tep_strerror(filter->tep, err, buf, buflen);
}
/**
@ -1443,7 +1440,7 @@ void tep_filter_reset(struct tep_event_filter *filter)
void tep_filter_free(struct tep_event_filter *filter)
{
tep_unref(filter->pevent);
tep_unref(filter->tep);
tep_filter_reset(filter);
@ -1462,10 +1459,10 @@ static int copy_filter_type(struct tep_event_filter *filter,
const char *name;
char *str;
/* Can't assume that the pevent's are the same */
/* Can't assume that the tep's are the same */
sys = filter_type->event->system;
name = filter_type->event->name;
event = tep_find_event_by_name(filter->pevent, sys, name);
event = tep_find_event_by_name(filter->tep, sys, name);
if (!event)
return -1;
@ -1522,167 +1519,6 @@ int tep_filter_copy(struct tep_event_filter *dest, struct tep_event_filter *sour
return ret;
}
/**
* tep_update_trivial - update the trivial filters with the given filter
* @dest - the filter to update
* @source - the filter as the source of the update
* @type - the type of trivial filter to update.
*
* Scan dest for trivial events matching @type to replace with the source.
*
* Returns 0 on success and -1 if there was a problem updating, but
* events may have still been updated on error.
*/
int tep_update_trivial(struct tep_event_filter *dest, struct tep_event_filter *source,
enum tep_filter_trivial_type type)
{
struct tep_handle *src_pevent;
struct tep_handle *dest_pevent;
struct tep_event *event;
struct tep_filter_type *filter_type;
struct tep_filter_arg *arg;
char *str;
int i;
src_pevent = source->pevent;
dest_pevent = dest->pevent;
/* Do nothing if either of the filters has nothing to filter */
if (!dest->filters || !source->filters)
return 0;
for (i = 0; i < dest->filters; i++) {
filter_type = &dest->event_filters[i];
arg = filter_type->filter;
if (arg->type != TEP_FILTER_ARG_BOOLEAN)
continue;
if ((arg->boolean.value && type == TEP_FILTER_TRIVIAL_FALSE) ||
(!arg->boolean.value && type == TEP_FILTER_TRIVIAL_TRUE))
continue;
event = filter_type->event;
if (src_pevent != dest_pevent) {
/* do a look up */
event = tep_find_event_by_name(src_pevent,
event->system,
event->name);
if (!event)
return -1;
}
str = tep_filter_make_string(source, event->id);
if (!str)
continue;
/* Don't bother if the filter is trivial too */
if (strcmp(str, "TRUE") != 0 && strcmp(str, "FALSE") != 0)
filter_event(dest, event, str, NULL);
free(str);
}
return 0;
}
/**
* tep_filter_clear_trivial - clear TRUE and FALSE filters
* @filter: the filter to remove trivial filters from
* @type: remove only true, false, or both
*
* Removes filters that only contain a TRUE or FALES boolean arg.
*
* Returns 0 on success and -1 if there was a problem.
*/
int tep_filter_clear_trivial(struct tep_event_filter *filter,
enum tep_filter_trivial_type type)
{
struct tep_filter_type *filter_type;
int count = 0;
int *ids = NULL;
int i;
if (!filter->filters)
return 0;
/*
* Two steps, first get all ids with trivial filters.
* then remove those ids.
*/
for (i = 0; i < filter->filters; i++) {
int *new_ids;
filter_type = &filter->event_filters[i];
if (filter_type->filter->type != TEP_FILTER_ARG_BOOLEAN)
continue;
switch (type) {
case TEP_FILTER_TRIVIAL_FALSE:
if (filter_type->filter->boolean.value)
continue;
break;
case TEP_FILTER_TRIVIAL_TRUE:
if (!filter_type->filter->boolean.value)
continue;
default:
break;
}
new_ids = realloc(ids, sizeof(*ids) * (count + 1));
if (!new_ids) {
free(ids);
return -1;
}
ids = new_ids;
ids[count++] = filter_type->event_id;
}
if (!count)
return 0;
for (i = 0; i < count; i++)
tep_filter_remove_event(filter, ids[i]);
free(ids);
return 0;
}
/**
* tep_filter_event_has_trivial - return true event contains trivial filter
* @filter: the filter with the information
* @event_id: the id of the event to test
* @type: trivial type to test for (TRUE, FALSE, EITHER)
*
* Returns 1 if the event contains a matching trivial type
* otherwise 0.
*/
int tep_filter_event_has_trivial(struct tep_event_filter *filter,
int event_id,
enum tep_filter_trivial_type type)
{
struct tep_filter_type *filter_type;
if (!filter->filters)
return 0;
filter_type = find_filter_type(filter, event_id);
if (!filter_type)
return 0;
if (filter_type->filter->type != TEP_FILTER_ARG_BOOLEAN)
return 0;
switch (type) {
case TEP_FILTER_TRIVIAL_FALSE:
return !filter_type->filter->boolean.value;
case TEP_FILTER_TRIVIAL_TRUE:
return filter_type->filter->boolean.value;
default:
return 1;
}
}
static int test_filter(struct tep_event *event, struct tep_filter_arg *arg,
struct tep_record *record, enum tep_errno *err);
@ -1692,8 +1528,8 @@ get_comm(struct tep_event *event, struct tep_record *record)
const char *comm;
int pid;
pid = tep_data_pid(event->pevent, record);
comm = tep_data_comm_from_pid(event->pevent, pid);
pid = tep_data_pid(event->tep, record);
comm = tep_data_comm_from_pid(event->tep, pid);
return comm;
}
@ -1861,7 +1697,7 @@ static int test_num(struct tep_event *event, struct tep_filter_arg *arg,
static const char *get_field_str(struct tep_filter_arg *arg, struct tep_record *record)
{
struct tep_event *event;
struct tep_handle *pevent;
struct tep_handle *tep;
unsigned long long addr;
const char *val = NULL;
unsigned int size;
@ -1891,12 +1727,12 @@ static const char *get_field_str(struct tep_filter_arg *arg, struct tep_record *
} else {
event = arg->str.field->event;
pevent = event->pevent;
tep = event->tep;
addr = get_value(event, arg->str.field, record);
if (arg->str.field->flags & (TEP_FIELD_IS_POINTER | TEP_FIELD_IS_LONG))
/* convert to a kernel symbol */
val = tep_find_function(pevent, addr);
val = tep_find_function(tep, addr);
if (val == NULL) {
/* just use the hex of the string name */
@ -2036,7 +1872,7 @@ int tep_event_filtered(struct tep_event_filter *filter, int event_id)
enum tep_errno tep_filter_match(struct tep_event_filter *filter,
struct tep_record *record)
{
struct tep_handle *pevent = filter->pevent;
struct tep_handle *tep = filter->tep;
struct tep_filter_type *filter_type;
int event_id;
int ret;
@ -2047,7 +1883,7 @@ enum tep_errno tep_filter_match(struct tep_event_filter *filter,
if (!filter->filters)
return TEP_ERRNO__NO_FILTER;
event_id = tep_data_type(pevent, record);
event_id = tep_data_type(tep, record);
filter_type = find_filter_type(filter, event_id);
if (!filter_type)
@ -2409,14 +2245,6 @@ int tep_filter_compare(struct tep_event_filter *filter1, struct tep_event_filter
break;
if (filter_type1->filter->type != filter_type2->filter->type)
break;
switch (filter_type1->filter->type) {
case TEP_FILTER_TRIVIAL_FALSE:
case TEP_FILTER_TRIVIAL_TRUE:
/* trivial types just need the type compared */
continue;
default:
break;
}
/* The best way to compare complex filters is with strings */
str1 = arg_to_str(filter1, filter_type1->filter);
str2 = arg_to_str(filter2, filter_type2->filter);

8
tools/lib/traceevent/plugin_cfg80211.c
View file

@ -25,9 +25,9 @@ process___le16_to_cpup(struct trace_seq *s, unsigned long long *args)
return val ? (long long) le16toh(*val) : 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_print_function(pevent,
tep_register_print_function(tep,
process___le16_to_cpup,
TEP_FUNC_ARG_INT,
"__le16_to_cpup",
@ -36,8 +36,8 @@ int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_print_function(pevent, process___le16_to_cpup,
tep_unregister_print_function(tep, process___le16_to_cpup,
"__le16_to_cpup");
}

14
tools/lib/traceevent/plugin_function.c
View file

@ -126,7 +126,7 @@ static int add_and_get_index(const char *parent, const char *child, int cpu)
static int function_handler(struct trace_seq *s, struct tep_record *record,
struct tep_event *event, void *context)
{
struct tep_handle *pevent = event->pevent;
struct tep_handle *tep = event->tep;
unsigned long long function;
unsigned long long pfunction;
const char *func;
@ -136,12 +136,12 @@ static int function_handler(struct trace_seq *s, struct tep_record *record,
if (tep_get_field_val(s, event, "ip", record, &function, 1))
return trace_seq_putc(s, '!');
func = tep_find_function(pevent, function);
func = tep_find_function(tep, function);
if (tep_get_field_val(s, event, "parent_ip", record, &pfunction, 1))
return trace_seq_putc(s, '!');
parent = tep_find_function(pevent, pfunction);
parent = tep_find_function(tep, pfunction);
if (parent && ftrace_indent->set)
index = add_and_get_index(parent, func, record->cpu);
@ -164,9 +164,9 @@ static int function_handler(struct trace_seq *s, struct tep_record *record,
return 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_event_handler(pevent, -1, "ftrace", "function",
tep_register_event_handler(tep, -1, "ftrace", "function",
function_handler, NULL);
tep_plugin_add_options("ftrace", plugin_options);
@ -174,11 +174,11 @@ int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
int i, x;
tep_unregister_event_handler(pevent, -1, "ftrace", "function",
tep_unregister_event_handler(tep, -1, "ftrace", "function",
function_handler, NULL);
for (i = 0; i <= cpus; i++) {

12
tools/lib/traceevent/plugin_hrtimer.c
View file

@ -67,23 +67,23 @@ static int timer_start_handler(struct trace_seq *s,
return 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_event_handler(pevent, -1,
tep_register_event_handler(tep, -1,
"timer", "hrtimer_expire_entry",
timer_expire_handler, NULL);
tep_register_event_handler(pevent, -1, "timer", "hrtimer_start",
tep_register_event_handler(tep, -1, "timer", "hrtimer_start",
timer_start_handler, NULL);
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_event_handler(pevent, -1,
tep_unregister_event_handler(tep, -1,
"timer", "hrtimer_expire_entry",
timer_expire_handler, NULL);
tep_unregister_event_handler(pevent, -1, "timer", "hrtimer_start",
tep_unregister_event_handler(tep, -1, "timer", "hrtimer_start",
timer_start_handler, NULL);
}

12
tools/lib/traceevent/plugin_jbd2.c
View file

@ -48,16 +48,16 @@ process_jiffies_to_msecs(struct trace_seq *s, unsigned long long *args)
return jiffies;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_print_function(pevent,
tep_register_print_function(tep,
process_jbd2_dev_to_name,
TEP_FUNC_ARG_STRING,
"jbd2_dev_to_name",
TEP_FUNC_ARG_INT,
TEP_FUNC_ARG_VOID);
tep_register_print_function(pevent,
tep_register_print_function(tep,
process_jiffies_to_msecs,
TEP_FUNC_ARG_LONG,
"jiffies_to_msecs",
@ -66,11 +66,11 @@ int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_print_function(pevent, process_jbd2_dev_to_name,
tep_unregister_print_function(tep, process_jbd2_dev_to_name,
"jbd2_dev_to_name");
tep_unregister_print_function(pevent, process_jiffies_to_msecs,
tep_unregister_print_function(tep, process_jiffies_to_msecs,
"jiffies_to_msecs");
}

32
tools/lib/traceevent/plugin_kmem.c
View file

@ -39,57 +39,57 @@ static int call_site_handler(struct trace_seq *s, struct tep_record *record,
if (tep_read_number_field(field, data, &val))
return 1;
func = tep_find_function(event->pevent, val);
func = tep_find_function(event->tep, val);
if (!func)
return 1;
addr = tep_find_function_address(event->pevent, val);
addr = tep_find_function_address(event->tep, val);
trace_seq_printf(s, "(%s+0x%x) ", func, (int)(val - addr));
return 1;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_event_handler(pevent, -1, "kmem", "kfree",
tep_register_event_handler(tep, -1, "kmem", "kfree",
call_site_handler, NULL);
tep_register_event_handler(pevent, -1, "kmem", "kmalloc",
tep_register_event_handler(tep, -1, "kmem", "kmalloc",
call_site_handler, NULL);
tep_register_event_handler(pevent, -1, "kmem", "kmalloc_node",
tep_register_event_handler(tep, -1, "kmem", "kmalloc_node",
call_site_handler, NULL);
tep_register_event_handler(pevent, -1, "kmem", "kmem_cache_alloc",
tep_register_event_handler(tep, -1, "kmem", "kmem_cache_alloc",
call_site_handler, NULL);
tep_register_event_handler(pevent, -1, "kmem",
tep_register_event_handler(tep, -1, "kmem",
"kmem_cache_alloc_node",
call_site_handler, NULL);
tep_register_event_handler(pevent, -1, "kmem", "kmem_cache_free",
tep_register_event_handler(tep, -1, "kmem", "kmem_cache_free",
call_site_handler, NULL);
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_event_handler(pevent, -1, "kmem", "kfree",
tep_unregister_event_handler(tep, -1, "kmem", "kfree",
call_site_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kmem", "kmalloc",
tep_unregister_event_handler(tep, -1, "kmem", "kmalloc",
call_site_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kmem", "kmalloc_node",
tep_unregister_event_handler(tep, -1, "kmem", "kmalloc_node",
call_site_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kmem", "kmem_cache_alloc",
tep_unregister_event_handler(tep, -1, "kmem", "kmem_cache_alloc",
call_site_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kmem",
tep_unregister_event_handler(tep, -1, "kmem",
"kmem_cache_alloc_node",
call_site_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kmem", "kmem_cache_free",
tep_unregister_event_handler(tep, -1, "kmem", "kmem_cache_free",
call_site_handler, NULL);
}

48
tools/lib/traceevent/plugin_kvm.c
View file

@ -389,8 +389,8 @@ static int kvm_mmu_print_role(struct trace_seq *s, struct tep_record *record,
* We can only use the structure if file is of the same
* endianness.
*/
if (tep_file_bigendian(event->pevent) ==
tep_is_host_bigendian(event->pevent)) {
if (tep_is_file_bigendian(event->tep) ==
tep_is_local_bigendian(event->tep)) {
trace_seq_printf(s, "%u q%u%s %s%s %spae %snxe %swp%s%s%s",
role.level,
@ -445,40 +445,40 @@ process_is_writable_pte(struct trace_seq *s, unsigned long long *args)
return pte & PT_WRITABLE_MASK;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
init_disassembler();
tep_register_event_handler(pevent, -1, "kvm", "kvm_exit",
tep_register_event_handler(tep, -1, "kvm", "kvm_exit",
kvm_exit_handler, NULL);
tep_register_event_handler(pevent, -1, "kvm", "kvm_emulate_insn",
tep_register_event_handler(tep, -1, "kvm", "kvm_emulate_insn",
kvm_emulate_insn_handler, NULL);
tep_register_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit",
tep_register_event_handler(tep, -1, "kvm", "kvm_nested_vmexit",
kvm_nested_vmexit_handler, NULL);
tep_register_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit_inject",
tep_register_event_handler(tep, -1, "kvm", "kvm_nested_vmexit_inject",
kvm_nested_vmexit_inject_handler, NULL);
tep_register_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_get_page",
tep_register_event_handler(tep, -1, "kvmmmu", "kvm_mmu_get_page",
kvm_mmu_get_page_handler, NULL);
tep_register_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_sync_page",
tep_register_event_handler(tep, -1, "kvmmmu", "kvm_mmu_sync_page",
kvm_mmu_print_role, NULL);
tep_register_event_handler(pevent, -1,
tep_register_event_handler(tep, -1,
"kvmmmu", "kvm_mmu_unsync_page",
kvm_mmu_print_role, NULL);
tep_register_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_zap_page",
tep_register_event_handler(tep, -1, "kvmmmu", "kvm_mmu_zap_page",
kvm_mmu_print_role, NULL);
tep_register_event_handler(pevent, -1, "kvmmmu",
tep_register_event_handler(tep, -1, "kvmmmu",
"kvm_mmu_prepare_zap_page", kvm_mmu_print_role,
NULL);
tep_register_print_function(pevent,
tep_register_print_function(tep,
process_is_writable_pte,
TEP_FUNC_ARG_INT,
"is_writable_pte",
@ -487,37 +487,37 @@ int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_event_handler(pevent, -1, "kvm", "kvm_exit",
tep_unregister_event_handler(tep, -1, "kvm", "kvm_exit",
kvm_exit_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kvm", "kvm_emulate_insn",
tep_unregister_event_handler(tep, -1, "kvm", "kvm_emulate_insn",
kvm_emulate_insn_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit",
tep_unregister_event_handler(tep, -1, "kvm", "kvm_nested_vmexit",
kvm_nested_vmexit_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kvm", "kvm_nested_vmexit_inject",
tep_unregister_event_handler(tep, -1, "kvm", "kvm_nested_vmexit_inject",
kvm_nested_vmexit_inject_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_get_page",
tep_unregister_event_handler(tep, -1, "kvmmmu", "kvm_mmu_get_page",
kvm_mmu_get_page_handler, NULL);
tep_unregister_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_sync_page",
tep_unregister_event_handler(tep, -1, "kvmmmu", "kvm_mmu_sync_page",
kvm_mmu_print_role, NULL);
tep_unregister_event_handler(pevent, -1,
tep_unregister_event_handler(tep, -1,
"kvmmmu", "kvm_mmu_unsync_page",
kvm_mmu_print_role, NULL);
tep_unregister_event_handler(pevent, -1, "kvmmmu", "kvm_mmu_zap_page",
tep_unregister_event_handler(tep, -1, "kvmmmu", "kvm_mmu_zap_page",
kvm_mmu_print_role, NULL);
tep_unregister_event_handler(pevent, -1, "kvmmmu",
tep_unregister_event_handler(tep, -1, "kvmmmu",
"kvm_mmu_prepare_zap_page", kvm_mmu_print_role,
NULL);
tep_unregister_print_function(pevent, process_is_writable_pte,
tep_unregister_print_function(tep, process_is_writable_pte,
"is_writable_pte");
}

8
tools/lib/traceevent/plugin_mac80211.c
View file

@ -87,17 +87,17 @@ static int drv_bss_info_changed(struct trace_seq *s,
return 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_event_handler(pevent, -1, "mac80211",
tep_register_event_handler(tep, -1, "mac80211",
"drv_bss_info_changed",
drv_bss_info_changed, NULL);
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_event_handler(pevent, -1, "mac80211",
tep_unregister_event_handler(tep, -1, "mac80211",
"drv_bss_info_changed",
drv_bss_info_changed, NULL);
}

18
tools/lib/traceevent/plugin_sched_switch.c
View file

@ -62,7 +62,7 @@ static void write_and_save_comm(struct tep_format_field *field,
comm = &s->buffer[len];
/* Help out the comm to ids. This will handle dups */
tep_register_comm(field->event->pevent, comm, pid);
tep_register_comm(field->event->tep, comm, pid);
}
static int sched_wakeup_handler(struct trace_seq *s,
@ -135,27 +135,27 @@ static int sched_switch_handler(struct trace_seq *s,
return 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_event_handler(pevent, -1, "sched", "sched_switch",
tep_register_event_handler(tep, -1, "sched", "sched_switch",
sched_switch_handler, NULL);
tep_register_event_handler(pevent, -1, "sched", "sched_wakeup",
tep_register_event_handler(tep, -1, "sched", "sched_wakeup",
sched_wakeup_handler, NULL);
tep_register_event_handler(pevent, -1, "sched", "sched_wakeup_new",
tep_register_event_handler(tep, -1, "sched", "sched_wakeup_new",
sched_wakeup_handler, NULL);
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_event_handler(pevent, -1, "sched", "sched_switch",
tep_unregister_event_handler(tep, -1, "sched", "sched_switch",
sched_switch_handler, NULL);
tep_unregister_event_handler(pevent, -1, "sched", "sched_wakeup",
tep_unregister_event_handler(tep, -1, "sched", "sched_wakeup",
sched_wakeup_handler, NULL);
tep_unregister_event_handler(pevent, -1, "sched", "sched_wakeup_new",
tep_unregister_event_handler(tep, -1, "sched", "sched_wakeup_new",
sched_wakeup_handler, NULL);
}

8
tools/lib/traceevent/plugin_scsi.c
View file

@ -414,9 +414,9 @@ unsigned long long process_scsi_trace_parse_cdb(struct trace_seq *s,
return 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_print_function(pevent,
tep_register_print_function(tep,
process_scsi_trace_parse_cdb,
TEP_FUNC_ARG_STRING,
"scsi_trace_parse_cdb",
@ -427,8 +427,8 @@ int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_print_function(pevent, process_scsi_trace_parse_cdb,
tep_unregister_print_function(tep, process_scsi_trace_parse_cdb,
"scsi_trace_parse_cdb");
}

8
tools/lib/traceevent/plugin_xen.c
View file

@ -120,9 +120,9 @@ unsigned long long process_xen_hypercall_name(struct trace_seq *s,
return 0;
}
int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
int TEP_PLUGIN_LOADER(struct tep_handle *tep)
{
tep_register_print_function(pevent,
tep_register_print_function(tep,
process_xen_hypercall_name,
TEP_FUNC_ARG_STRING,
"xen_hypercall_name",
@ -131,8 +131,8 @@ int TEP_PLUGIN_LOADER(struct tep_handle *pevent)
return 0;
}
void TEP_PLUGIN_UNLOADER(struct tep_handle *pevent)
void TEP_PLUGIN_UNLOADER(struct tep_handle *tep)
{
tep_unregister_print_function(pevent, process_xen_hypercall_name,
tep_unregister_print_function(tep, process_xen_hypercall_name,
"xen_hypercall_name");
}

19
tools/perf/Documentation/perf-record.txt
View file

@ -459,6 +459,25 @@ Set affinity mask of trace reading thread according to the policy defined by 'mo
node - thread affinity mask is set to NUMA node cpu mask of the processed mmap buffer
cpu - thread affinity mask is set to cpu of the processed mmap buffer
--mmap-flush=number::
Specify minimal number of bytes that is extracted from mmap data pages and
processed for output. One can specify the number using B/K/M/G suffixes.
The maximal allowed value is a quarter of the size of mmaped data pages.
The default option value is 1 byte which means that every time that the output
writing thread finds some new data in the mmaped buffer the data is extracted,
possibly compressed (-z) and written to the output, perf.data or pipe.
Larger data chunks are compressed more effectively in comparison to smaller
chunks so extraction of larger chunks from the mmap data pages is preferable
from the perspective of output size reduction.
Also at some cases executing less output write syscalls with bigger data size
can take less time than executing more output write syscalls with smaller data
size thus lowering runtime profiling overhead.
--all-kernel::
Configure all used events to run in kernel space.

20
tools/perf/Makefile.config
View file

@ -152,6 +152,13 @@ endif
FEATURE_CHECK_CFLAGS-libbabeltrace := $(LIBBABELTRACE_CFLAGS)
FEATURE_CHECK_LDFLAGS-libbabeltrace := $(LIBBABELTRACE_LDFLAGS) -lbabeltrace-ctf
ifdef LIBZSTD_DIR
LIBZSTD_CFLAGS := -I$(LIBZSTD_DIR)/lib
LIBZSTD_LDFLAGS := -L$(LIBZSTD_DIR)/lib
endif
FEATURE_CHECK_CFLAGS-libzstd := $(LIBZSTD_CFLAGS)
FEATURE_CHECK_LDFLAGS-libzstd := $(LIBZSTD_LDFLAGS)
FEATURE_CHECK_CFLAGS-bpf = -I. -I$(srctree)/tools/include -I$(srctree)/tools/arch/$(SRCARCH)/include/uapi -I$(srctree)/tools/include/uapi
# include ARCH specific config
-include $(src-perf)/arch/$(SRCARCH)/Makefile
@ -787,6 +794,19 @@ ifndef NO_LZMA
endif
endif
ifndef NO_LIBZSTD
ifeq ($(feature-libzstd), 1)
CFLAGS += -DHAVE_ZSTD_SUPPORT
CFLAGS += $(LIBZSTD_CFLAGS)
LDFLAGS += $(LIBZSTD_LDFLAGS)
EXTLIBS += -lzstd
$(call detected,CONFIG_ZSTD)
else
msg := $(warning No libzstd found, disables trace compression, please install libzstd-dev[el] and/or set LIBZSTD_DIR);
NO_LIBZSTD := 1
endif
endif
ifndef NO_BACKTRACE
ifeq ($(feature-backtrace), 1)
CFLAGS += -DHAVE_BACKTRACE_SUPPORT

3
tools/perf/Makefile.perf
View file

@ -108,6 +108,9 @@ include ../scripts/utilities.mak
# streaming for record mode. Currently Posix AIO trace streaming is
# supported only when linking with glibc.
#
# Define NO_LIBZSTD if you do not want support of Zstandard based runtime
# trace compression in record mode.
#
# As per kernel Makefile, avoid funny character set dependencies
unexport LC_ALL

2
tools/perf/builtin-kmem.c
View file

@ -1975,7 +1975,7 @@ int cmd_kmem(int argc, const char **argv)
goto out_delete;
}
kmem_page_size = tep_get_page_size(evsel->tp_format->pevent);
kmem_page_size = tep_get_page_size(evsel->tp_format->tep);
symbol_conf.use_callchain = true;
}

6
tools/perf/builtin-list.c
View file

@ -70,10 +70,11 @@ int cmd_list(int argc, const char **argv)
print_symbol_events(NULL, PERF_TYPE_HARDWARE,
event_symbols_hw, PERF_COUNT_HW_MAX, raw_dump);
else if (strcmp(argv[i], "sw") == 0 ||
strcmp(argv[i], "software") == 0)
strcmp(argv[i], "software") == 0) {
print_symbol_events(NULL, PERF_TYPE_SOFTWARE,
event_symbols_sw, PERF_COUNT_SW_MAX, raw_dump);
else if (strcmp(argv[i], "cache") == 0 ||
print_tool_events(NULL, raw_dump);
} else if (strcmp(argv[i], "cache") == 0 ||
strcmp(argv[i], "hwcache") == 0)
print_hwcache_events(NULL, raw_dump);
else if (strcmp(argv[i], "pmu") == 0)
@ -113,6 +114,7 @@ int cmd_list(int argc, const char **argv)
event_symbols_hw, PERF_COUNT_HW_MAX, raw_dump);
print_symbol_events(s, PERF_TYPE_SOFTWARE,
event_symbols_sw, PERF_COUNT_SW_MAX, raw_dump);
print_tool_events(s, raw_dump);
print_hwcache_events(s, raw_dump);
print_pmu_events(s, raw_dump, !desc_flag,
long_desc_flag,

65
tools/perf/builtin-record.c
View file

@ -337,6 +337,41 @@ static int record__aio_enabled(struct record *rec)
return rec->opts.nr_cblocks > 0;
}
#define MMAP_FLUSH_DEFAULT 1
static int record__mmap_flush_parse(const struct option *opt,
const char *str,
int unset)
{
int flush_max;
struct record_opts *opts = (struct record_opts *)opt->value;
static struct parse_tag tags[] = {
{ .tag = 'B', .mult = 1 },
{ .tag = 'K', .mult = 1 << 10 },
{ .tag = 'M', .mult = 1 << 20 },
{ .tag = 'G', .mult = 1 << 30 },
{ .tag = 0 },
};
if (unset)
return 0;
if (str) {
opts->mmap_flush = parse_tag_value(str, tags);
if (opts->mmap_flush == (int)-1)
opts->mmap_flush = strtol(str, NULL, 0);
}
if (!opts->mmap_flush)
opts->mmap_flush = MMAP_FLUSH_DEFAULT;
flush_max = perf_evlist__mmap_size(opts->mmap_pages);
flush_max /= 4;
if (opts->mmap_flush > flush_max)
opts->mmap_flush = flush_max;
return 0;
}
static int process_synthesized_event(struct perf_tool *tool,
union perf_event *event,
struct perf_sample *sample __maybe_unused,
@ -546,7 +581,8 @@ static int record__mmap_evlist(struct record *rec,
if (perf_evlist__mmap_ex(evlist, opts->mmap_pages,
opts->auxtrace_mmap_pages,
opts->auxtrace_snapshot_mode,
opts->nr_cblocks, opts->affinity) < 0) {
opts->nr_cblocks, opts->affinity,
opts->mmap_flush) < 0) {
if (errno == EPERM) {
pr_err("Permission error mapping pages.\n"
"Consider increasing "
@ -736,7 +772,7 @@ static void record__adjust_affinity(struct record *rec, struct perf_mmap *map)
}
static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evlist,
bool overwrite)
bool overwrite, bool synch)
{
u64 bytes_written = rec->bytes_written;
int i;
@ -759,12 +795,19 @@ static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evli
off = record__aio_get_pos(trace_fd);
for (i = 0; i < evlist->nr_mmaps; i++) {
u64 flush = 0;
struct perf_mmap *map = &maps[i];
if (map->base) {
record__adjust_affinity(rec, map);
if (synch) {
flush = map->flush;
map->flush = 1;
}
if (!record__aio_enabled(rec)) {
if (perf_mmap__push(map, rec, record__pushfn) != 0) {
if (synch)
map->flush = flush;
rc = -1;
goto out;
}
@ -777,10 +820,14 @@ static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evli
idx = record__aio_sync(map, false);
if (perf_mmap__aio_push(map, rec, idx, record__aio_pushfn, &off) != 0) {
record__aio_set_pos(trace_fd, off);
if (synch)
map->flush = flush;
rc = -1;
goto out;
}
}
if (synch)
map->flush = flush;
}
if (map->auxtrace_mmap.base && !rec->opts.auxtrace_snapshot_mode &&
@ -806,15 +853,15 @@ static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evli
return rc;
}
static int record__mmap_read_all(struct record *rec)
static int record__mmap_read_all(struct record *rec, bool synch)
{
int err;
err = record__mmap_read_evlist(rec, rec->evlist, false);
err = record__mmap_read_evlist(rec, rec->evlist, false, synch);
if (err)
return err;
return record__mmap_read_evlist(rec, rec->evlist, true);
return record__mmap_read_evlist(rec, rec->evlist, true, synch);
}
static void record__init_features(struct record *rec)
@ -1340,7 +1387,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
if (trigger_is_hit(&switch_output_trigger) || done || draining)
perf_evlist__toggle_bkw_mmap(rec->evlist, BKW_MMAP_DATA_PENDING);
if (record__mmap_read_all(rec) < 0) {
if (record__mmap_read_all(rec, false) < 0) {
trigger_error(&auxtrace_snapshot_trigger);
trigger_error(&switch_output_trigger);
err = -1;
@ -1441,6 +1488,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
record__synthesize_workload(rec, true);
out_child:
record__mmap_read_all(rec, true);
record__aio_mmap_read_sync(rec);
if (forks) {
@ -1846,6 +1894,7 @@ static struct record record = {
.uses_mmap = true,
.default_per_cpu = true,
},
.mmap_flush = MMAP_FLUSH_DEFAULT,
},
.tool = {
.sample = process_sample_event,
@ -1912,6 +1961,9 @@ static struct option __record_options[] = {
OPT_CALLBACK('m', "mmap-pages", &record.opts, "pages[,pages]",
"number of mmap data pages and AUX area tracing mmap pages",
record__parse_mmap_pages),
OPT_CALLBACK(0, "mmap-flush", &record.opts, "number",
"Minimal number of bytes that is extracted from mmap data pages (default: 1)",
record__mmap_flush_parse),
OPT_BOOLEAN(0, "group", &record.opts.group,
"put the counters into a counter group"),
OPT_CALLBACK_NOOPT('g', NULL, &callchain_param,
@ -2224,6 +2276,7 @@ int cmd_record(int argc, const char **argv)
pr_info("nr_cblocks: %d\n", rec->opts.nr_cblocks);
pr_debug("affinity: %s\n", affinity_tags[rec->opts.affinity]);
pr_debug("mmap flush: %d\n", rec->opts.mmap_flush);
err = __cmd_record(&record, argc, argv);
out:

28
tools/perf/builtin-stat.c
View file

@ -244,11 +244,25 @@ perf_evsel__write_stat_event(struct perf_evsel *counter, u32 cpu, u32 thread,
process_synthesized_event, NULL);
}
static int read_single_counter(struct perf_evsel *counter, int cpu,
int thread, struct timespec *rs)
{
if (counter->tool_event == PERF_TOOL_DURATION_TIME) {
u64 val = rs->tv_nsec + rs->tv_sec*1000000000ULL;
struct perf_counts_values *count =
perf_counts(counter->counts, cpu, thread);
count->ena = count->run = val;
count->val = val;
return 0;
}
return perf_evsel__read_counter(counter, cpu, thread);
}
/*
* Read out the results of a single counter:
* do not aggregate counts across CPUs in system-wide mode
*/
static int read_counter(struct perf_evsel *counter)
static int read_counter(struct perf_evsel *counter, struct timespec *rs)
{
int nthreads = thread_map__nr(evsel_list->threads);
int ncpus, cpu, thread;
@ -275,7 +289,7 @@ static int read_counter(struct perf_evsel *counter)
* (via perf_evsel__read_counter) and sets threir count->loaded.
*/
if (!count->loaded &&
perf_evsel__read_counter(counter, cpu, thread)) {
read_single_counter(counter, cpu, thread, rs)) {
counter->counts->scaled = -1;
perf_counts(counter->counts, cpu, thread)->ena = 0;
perf_counts(counter->counts, cpu, thread)->run = 0;
@ -304,13 +318,13 @@ static int read_counter(struct perf_evsel *counter)
return 0;
}
static void read_counters(void)
static void read_counters(struct timespec *rs)
{
struct perf_evsel *counter;
int ret;
evlist__for_each_entry(evsel_list, counter) {
ret = read_counter(counter);
ret = read_counter(counter, rs);
if (ret)
pr_debug("failed to read counter %s\n", counter->name);
@ -323,11 +337,11 @@ static void process_interval(void)
{
struct timespec ts, rs;
read_counters();
clock_gettime(CLOCK_MONOTONIC, &ts);
diff_timespec(&rs, &ts, &ref_time);
read_counters(&rs);
if (STAT_RECORD) {
if (WRITE_STAT_ROUND_EVENT(rs.tv_sec * NSEC_PER_SEC + rs.tv_nsec, INTERVAL))
pr_err("failed to write stat round event\n");
@ -593,7 +607,7 @@ static int __run_perf_stat(int argc, const char **argv, int run_idx)
* avoid arbitrary skew, we must read all counters before closing any
* group leaders.
*/
read_counters();
read_counters(&(struct timespec) { .tv_nsec = t1-t0 });
perf_evlist__close(evsel_list);
return WEXITSTATUS(status);

2
tools/perf/builtin-version.c
View file

@ -78,6 +78,8 @@ static void library_status(void)
STATUS(HAVE_LZMA_SUPPORT, lzma);
STATUS(HAVE_AUXTRACE_SUPPORT, get_cpuid);
STATUS(HAVE_LIBBPF_SUPPORT, bpf);
STATUS(HAVE_AIO_SUPPORT, aio);
STATUS(HAVE_ZSTD_SUPPORT, zstd);
}
int cmd_version(int argc, const char **argv)

196
tools/perf/examples/bpf/augmented_raw_syscalls.c
View file

@ -15,6 +15,7 @@
*/
#include <unistd.h>
#include <linux/limits.h>
#include <pid_filter.h>
/* bpf-output associated map */
@ -41,32 +42,110 @@ struct syscall_exit_args {
struct augmented_filename {
unsigned int size;
int reserved;
char value[256];
char value[PATH_MAX];
};
#define SYS_OPEN 2
#define SYS_ACCESS 21
#define SYS_OPENAT 257
/* syscalls where the first arg is a string */
#define SYS_OPEN 2
#define SYS_STAT 4
#define SYS_LSTAT 6
#define SYS_ACCESS 21
#define SYS_EXECVE 59
#define SYS_TRUNCATE 76
#define SYS_CHDIR 80
#define SYS_RENAME 82
#define SYS_MKDIR 83
#define SYS_RMDIR 84
#define SYS_CREAT 85
#define SYS_LINK 86
#define SYS_UNLINK 87
#define SYS_SYMLINK 88
#define SYS_READLINK 89
#define SYS_CHMOD 90
#define SYS_CHOWN 92
#define SYS_LCHOWN 94
#define SYS_MKNOD 133
#define SYS_STATFS 137
#define SYS_PIVOT_ROOT 155
#define SYS_CHROOT 161
#define SYS_ACCT 163
#define SYS_SWAPON 167
#define SYS_SWAPOFF 168
#define SYS_DELETE_MODULE 176
#define SYS_SETXATTR 188
#define SYS_LSETXATTR 189
#define SYS_GETXATTR 191
#define SYS_LGETXATTR 192
#define SYS_LISTXATTR 194
#define SYS_LLISTXATTR 195
#define SYS_REMOVEXATTR 197
#define SYS_LREMOVEXATTR 198
#define SYS_MQ_OPEN 240
#define SYS_MQ_UNLINK 241
#define SYS_ADD_KEY 248
#define SYS_REQUEST_KEY 249
#define SYS_SYMLINKAT 266
#define SYS_MEMFD_CREATE 319
/* syscalls where the first arg is a string */
#define SYS_PWRITE64 18
#define SYS_EXECVE 59
#define SYS_RENAME 82
#define SYS_QUOTACTL 179
#define SYS_FSETXATTR 190
#define SYS_FGETXATTR 193
#define SYS_FREMOVEXATTR 199
#define SYS_MQ_TIMEDSEND 242
#define SYS_REQUEST_KEY 249
#define SYS_INOTIFY_ADD_WATCH 254
#define SYS_OPENAT 257
#define SYS_MKDIRAT 258
#define SYS_MKNODAT 259
#define SYS_FCHOWNAT 260
#define SYS_FUTIMESAT 261
#define SYS_NEWFSTATAT 262
#define SYS_UNLINKAT 263
#define SYS_RENAMEAT 264
#define SYS_LINKAT 265
#define SYS_READLINKAT 267
#define SYS_FCHMODAT 268
#define SYS_FACCESSAT 269
#define SYS_UTIMENSAT 280
#define SYS_NAME_TO_HANDLE_AT 303
#define SYS_FINIT_MODULE 313
#define SYS_RENAMEAT2 316
#define SYS_EXECVEAT 322
#define SYS_STATX 332
pid_filter(pids_filtered);
struct augmented_args_filename {
struct syscall_enter_args args;
struct augmented_filename filename;
};
bpf_map(augmented_filename_map, PERCPU_ARRAY, int, struct augmented_args_filename, 1);
SEC("raw_syscalls:sys_enter")
int sys_enter(struct syscall_enter_args *args)
{
struct {
struct syscall_enter_args args;
struct augmented_filename filename;
} augmented_args;
struct syscall *syscall;
unsigned int len = sizeof(augmented_args);
struct augmented_args_filename *augmented_args;
unsigned int len = sizeof(*augmented_args);
const void *filename_arg = NULL;
struct syscall *syscall;
int key = 0;
augmented_args = bpf_map_lookup_elem(&augmented_filename_map, &key);
if (augmented_args == NULL)
return 1;
if (pid_filter__has(&pids_filtered, getpid()))
return 0;
probe_read(&augmented_args.args, sizeof(augmented_args.args), args);
probe_read(&augmented_args->args, sizeof(augmented_args->args), args);
syscall = bpf_map_lookup_elem(&syscalls, &augmented_args.args.syscall_nr);
syscall = bpf_map_lookup_elem(&syscalls, &augmented_args->args.syscall_nr);
if (syscall == NULL || !syscall->enabled)
return 0;
/*
@ -109,30 +188,105 @@ int sys_enter(struct syscall_enter_args *args)
*
* after the ctx memory access to prevent their down stream merging.
*/
switch (augmented_args.args.syscall_nr) {
/*
* This table of what args are strings will be provided by userspace,
* in the syscalls map, i.e. we will already have to do the lookup to
* see if this specific syscall is filtered, so we can as well get more
* info about what syscall args are strings or pointers, and how many
* bytes to copy, per arg, etc.
*
* For now hard code it, till we have all the basic mechanisms in place
* to automate everything and make the kernel part be completely driven
* by information obtained in userspace for each kernel version and
* processor architecture, making the kernel part the same no matter what
* kernel version or processor architecture it runs on.
*/
switch (augmented_args->args.syscall_nr) {
case SYS_ACCT:
case SYS_ADD_KEY:
case SYS_CHDIR:
case SYS_CHMOD:
case SYS_CHOWN:
case SYS_CHROOT:
case SYS_CREAT:
case SYS_DELETE_MODULE:
case SYS_EXECVE:
case SYS_GETXATTR:
case SYS_LCHOWN:
case SYS_LGETXATTR:
case SYS_LINK:
case SYS_LISTXATTR:
case SYS_LLISTXATTR:
case SYS_LREMOVEXATTR:
case SYS_LSETXATTR:
case SYS_LSTAT:
case SYS_MEMFD_CREATE:
case SYS_MKDIR:
case SYS_MKNOD:
case SYS_MQ_OPEN:
case SYS_MQ_UNLINK:
case SYS_PIVOT_ROOT:
case SYS_READLINK:
case SYS_REMOVEXATTR:
case SYS_RENAME:
case SYS_REQUEST_KEY:
case SYS_RMDIR:
case SYS_SETXATTR:
case SYS_STAT:
case SYS_STATFS:
case SYS_SWAPOFF:
case SYS_SWAPON:
case SYS_SYMLINK:
case SYS_SYMLINKAT:
case SYS_TRUNCATE:
case SYS_UNLINK:
case SYS_ACCESS:
case SYS_OPEN: filename_arg = (const void *)args->args[0];
__asm__ __volatile__("": : :"memory");
break;
case SYS_EXECVEAT:
case SYS_FACCESSAT:
case SYS_FCHMODAT:
case SYS_FCHOWNAT:
case SYS_FGETXATTR:
case SYS_FINIT_MODULE:
case SYS_FREMOVEXATTR:
case SYS_FSETXATTR:
case SYS_FUTIMESAT:
case SYS_INOTIFY_ADD_WATCH:
case SYS_LINKAT:
case SYS_MKDIRAT:
case SYS_MKNODAT:
case SYS_MQ_TIMEDSEND:
case SYS_NAME_TO_HANDLE_AT:
case SYS_NEWFSTATAT:
case SYS_PWRITE64:
case SYS_QUOTACTL:
case SYS_READLINKAT:
case SYS_RENAMEAT:
case SYS_RENAMEAT2:
case SYS_STATX:
case SYS_UNLINKAT:
case SYS_UTIMENSAT:
case SYS_OPENAT: filename_arg = (const void *)args->args[1];
break;
}
if (filename_arg != NULL) {
augmented_args.filename.reserved = 0;
augmented_args.filename.size = probe_read_str(&augmented_args.filename.value,
sizeof(augmented_args.filename.value),
augmented_args->filename.reserved = 0;
augmented_args->filename.size = probe_read_str(&augmented_args->filename.value,
sizeof(augmented_args->filename.value),
filename_arg);
if (augmented_args.filename.size < sizeof(augmented_args.filename.value)) {
len -= sizeof(augmented_args.filename.value) - augmented_args.filename.size;
len &= sizeof(augmented_args.filename.value) - 1;
if (augmented_args->filename.size < sizeof(augmented_args->filename.value)) {
len -= sizeof(augmented_args->filename.value) - augmented_args->filename.size;
len &= sizeof(augmented_args->filename.value) - 1;
}
} else {
len = sizeof(augmented_args.args);
len = sizeof(augmented_args->args);
}
/* If perf_event_output fails, return non-zero so that it gets recorded unaugmented */
return perf_event_output(args, &__augmented_syscalls__, BPF_F_CURRENT_CPU, &augmented_args, len);
return perf_event_output(args, &__augmented_syscalls__, BPF_F_CURRENT_CPU, augmented_args, len);
}
SEC("raw_syscalls:sys_exit")

1
tools/perf/perf.h
View file

@ -85,6 +85,7 @@ struct record_opts {
u64 clockid_res_ns;
int nr_cblocks;
int affinity;
int mmap_flush;
};
enum perf_affinity {

2
tools/perf/pmu-events/arch/s390/cf_z14/extended.json
View file

@ -4,7 +4,7 @@
"EventCode": "128",
"EventName": "L1D_RO_EXCL_WRITES",
"BriefDescription": "L1D Read-only Exclusive Writes",
"PublicDescription": "Counter:128 Name:L1D_RO_EXCL_WRITES A directory write to the Level-1 Data cache where the line was originally in a Read-Only state in the cache but has been updated to be in the Exclusive state that allows stores to the cache line"
"PublicDescription": "L1D_RO_EXCL_WRITES A directory write to the Level-1 Data cache where the line was originally in a Read-Only state in the cache but has been updated to be in the Exclusive state that allows stores to the cache line"
},
{
"Unit": "CPU-M-CF",

2
tools/perf/pmu-events/arch/x86/bonnell/frontend.json
View file

@ -77,7 +77,7 @@
"UMask": "0x1",
"EventName": "UOPS.MS_CYCLES",
"SampleAfterValue": "2000000",
"BriefDescription": "This event counts the cycles where 1 or more uops are issued by the micro-sequencer (MS), including microcode assists and inserted flows, and written to the IQ. ",
"BriefDescription": "This event counts the cycles where 1 or more uops are issued by the micro-sequencer (MS), including microcode assists and inserted flows, and written to the IQ.",
"CounterMask": "1"
}
]

2
tools/perf/pmu-events/arch/x86/bonnell/pipeline.json
View file

@ -189,7 +189,7 @@
"UMask": "0x8",
"EventName": "BR_MISSP_TYPE_RETIRED.IND_CALL",
"SampleAfterValue": "200000",
"BriefDescription": "Mispredicted indirect calls, including both register and memory indirect. "
"BriefDescription": "Mispredicted indirect calls, including both register and memory indirect."
},
{
"EventCode": "0x89",

260
tools/perf/pmu-events/arch/x86/broadwell/bdw-metrics.json
View file

@ -1,164 +1,352 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , IDQ.MITE_UOPS / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 16 * ( ICACHE.HIT + ICACHE.MISSES ) / 4.0 ) )",
"MetricGroup": "Frontend",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE.IFDATA_STALL - (( 14 * ITLB_MISSES.STLB_HIT + cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + 7* ITLB_MISSES.WALK_COMPLETED )) ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) * (12 * ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) ) * (4 * cycles) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "Branch_Misprediction_Cost"
},
{
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) * (12 * ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) ) * (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts_SMT",
"MetricName": "Branch_Misprediction_Cost_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_UOPS_RETIRED.L1_MISS + mem_load_uops_retired.hit_lfb )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( cpu@l1d_pend_miss.pending_cycles\\,any\\=1@ / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + cpu@DTLB_LOAD_MISSES.WALK_DURATION\\,cmask\\=1@ + cpu@DTLB_STORE_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * ( DTLB_STORE_MISSES.WALK_COMPLETED + DTLB_LOAD_MISSES.WALK_COMPLETED + ITLB_MISSES.WALK_COMPLETED ) ) / cycles",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + cpu@DTLB_LOAD_MISSES.WALK_DURATION\\,cmask\\=1@ + cpu@DTLB_STORE_MISSES.WALK_DURATION\\,cmask\\=1@ + 7*(DTLB_STORE_MISSES.WALK_COMPLETED+DTLB_LOAD_MISSES.WALK_COMPLETED+ITLB_MISSES.WALK_COMPLETED)) / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + cpu@DTLB_LOAD_MISSES.WALK_DURATION\\,cmask\\=1@ + cpu@DTLB_STORE_MISSES.WALK_DURATION\\,cmask\\=1@ + 7 * ( DTLB_STORE_MISSES.WALK_COMPLETED + DTLB_LOAD_MISSES.WALK_COMPLETED + ITLB_MISSES.WALK_COMPLETED ) ) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * L2_RQSTS.MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * ( L2_RQSTS.REFERENCES - L2_RQSTS.MISS ) / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2* FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4*( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8* FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "64 * ( arb@event\\=0x81\\,umask\\=0x1@ + arb@event\\=0x84\\,umask\\=0x1@ ) / 1000000 / duration_time / 1000",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

1630
tools/perf/pmu-events/arch/x86/broadwell/cache.json
View file

File diff suppressed because it is too large Load diff

51
tools/perf/pmu-events/arch/x86/broadwell/floating-point.json
View file

@ -1,24 +1,26 @@
[
{
"PublicDescription": "This event counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x8",
"Errata": "BDM30",
"EventName": "OTHER_ASSISTS.AVX_TO_SSE",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.",
"BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x10",
"Errata": "BDM30",
"EventName": "OTHER_ASSISTS.SSE_TO_AVX",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.",
"BriefDescription": "Number of transitions from legacy SSE to AVX-256 when penalty applicable (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -45,7 +47,7 @@
"UMask": "0x3",
"EventName": "FP_ARITH_INST_RETIRED.SCALAR",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. (RSQRT for single precision?)",
"CounterHTOff": "0,1,2,3"
},
{
@ -54,7 +56,7 @@
"UMask": "0x4",
"EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired. Each count represents 2 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired. Each count represents 2 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"CounterHTOff": "0,1,2,3"
},
{
@ -63,7 +65,7 @@
"UMask": "0x8",
"EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"CounterHTOff": "0,1,2,3"
},
{
@ -72,7 +74,7 @@
"UMask": "0x10",
"EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"CounterHTOff": "0,1,2,3"
},
{
@ -81,7 +83,7 @@
"UMask": "0x15",
"EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
"SampleAfterValue": "2000006",
"BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
"BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"CounterHTOff": "0,1,2,3"
},
{
@ -90,7 +92,7 @@
"UMask": "0x20",
"EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"CounterHTOff": "0,1,2,3"
},
{
@ -99,7 +101,7 @@
"UMask": "0x2a",
"EventName": "FP_ARITH_INST_RETIRED.SINGLE",
"SampleAfterValue": "2000005",
"BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
"BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"CounterHTOff": "0,1,2,3"
},
{
@ -108,57 +110,62 @@
"UMask": "0x3c",
"EventName": "FP_ARITH_INST_RETIRED.PACKED",
"SampleAfterValue": "2000004",
"BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. (RSQRT for single-precision?)",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "FP_ASSIST.X87_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to output value.",
"BriefDescription": "output - Numeric Overflow, Numeric Underflow, Inexact Result (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "FP_ASSIST.X87_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to input value.",
"BriefDescription": "input - Invalid Operation, Denormal Operand, SNaN Operand (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "FP_ASSIST.SIMD_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to Output values",
"BriefDescription": "SSE* FP micro-code assist when output value is invalid. (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts any input SSE* floating-point (FP) assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "FP_ASSIST.SIMD_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to input values",
"BriefDescription": "Any input SSE* FP Assist - (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts cycles with any input and output SSE or x87 FP assist. If an input and output assist are detected on the same cycle the event increments by 1.",
"PEBS": "1",
"PublicDescription": "This event counts cycles with any input and output SSE or x87 FP assist. If an input and output assist are detected on the same cycle the event increments by 1. Uses PEBS.",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x1e",
"EventName": "FP_ASSIST.ANY",
"SampleAfterValue": "100003",
"BriefDescription": "Cycles with any input/output SSE or FP assist",
"BriefDescription": "Counts any FP_ASSIST umask was incrementing (Precise Event)",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
}

4
tools/perf/pmu-events/arch/x86/broadwell/frontend.json
View file

@ -211,7 +211,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
"PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding \u201c4 \u2013 x\u201d when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
"EventCode": "0x9C",
"Counter": "0,1,2,3",
"UMask": "0x1",
@ -274,7 +274,7 @@
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
"PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 0\u20132 cycles.",
"EventCode": "0xAB",
"Counter": "0,1,2,3",
"UMask": "0x2",

1640
tools/perf/pmu-events/arch/x86/broadwell/memory.json
View file

File diff suppressed because it is too large Load diff

36
tools/perf/pmu-events/arch/x86/broadwell/pipeline.json
View file

@ -1,7 +1,6 @@
[
{
"PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. \nNotes: INST_RETIRED.ANY is counted by a designated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. INST_RETIRED.ANY_P is counted by a programmable counter and it is an architectural performance event. \nCounting: Faulting executions of GETSEC/VM entry/VM Exit/MWait will not count as retired instructions.",
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -11,7 +10,6 @@
},
{
"PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -20,7 +18,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"AnyThread": "1",
@ -31,7 +28,6 @@
},
{
"PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. \nNote: On all current platforms this event stops counting during 'throttling (TM)' states duty off periods the processor is 'halted'. This event is clocked by base clock (100 Mhz) on Sandy Bridge. The counter update is done at a lower clock rate then the core clock the overflow status bit for this counter may appear 'sticky'. After the counter has overflowed and software clears the overflow status bit and resets the counter to less than MAX. The reset value to the counter is not clocked immediately so the overflow status bit will flip 'high (1)' and generate another PMI (if enabled) after which the reset value gets clocked into the counter. Therefore, software will get the interrupt, read the overflow status bit '1 for bit 34 while the counter value is less than MAX. Software should ignore this case.",
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",
@ -317,7 +313,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts stalls occurred due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
"PublicDescription": "This event counts stalls occured due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
"EventCode": "0x87",
"Counter": "0,1,2,3",
"UMask": "0x1",
@ -786,8 +782,8 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
"EventCode": "0xA2",
"PublicDescription": "This event counts resource-related stall cycles.",
"EventCode": "0xa2",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "RESOURCE_STALLS.ANY",
@ -973,6 +969,7 @@
"CounterHTOff": "2"
},
{
"PublicDescription": "Number of Uops delivered by the LSD.",
"EventCode": "0xA8",
"Counter": "0,1,2,3",
"UMask": "0x1",
@ -1147,7 +1144,8 @@
"CounterHTOff": "1"
},
{
"PublicDescription": "This event counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
"EventCode": "0xC0",
"Counter": "0,1,2,3",
"UMask": "0x2",
@ -1157,12 +1155,12 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PEBS": "1",
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x40",
"EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
"SampleAfterValue": "100003",
"BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -1178,26 +1176,28 @@
"Data_LA": "1"
},
{
"PublicDescription": "This event counts cycles without actually retired uops.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts cycles without actually retired uops.",
"EventCode": "0xC2",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "UOPS_RETIRED.STALL_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles without actually retired uops.",
"BriefDescription": "Cycles no executable uops retired (Precise Event)",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
"PEBS": "1",
"PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to PEBS uops retired event.",
"EventCode": "0xC2",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "UOPS_RETIRED.TOTAL_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with less than 10 actually retired uops.",
"BriefDescription": "Number of cycles using always true condition applied to PEBS uops retired event.",
"CounterMask": "10",
"CounterHTOff": "0,1,2,3"
},
@ -1320,13 +1320,14 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts not taken branch instructions retired.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts not taken branch instructions retired.",
"EventCode": "0xC4",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "BR_INST_RETIRED.NOT_TAKEN",
"SampleAfterValue": "400009",
"BriefDescription": "Not taken branch instructions retired.",
"BriefDescription": "Counts all not taken macro branch instructions retired. (Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -1341,14 +1342,15 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts far branch instructions retired.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts far branch instructions retired.",
"EventCode": "0xC4",
"Counter": "0,1,2,3",
"UMask": "0x40",
"Errata": "BDW98",
"EventName": "BR_INST_RETIRED.FAR_BRANCH",
"SampleAfterValue": "100007",
"BriefDescription": "Far branch instructions retired.",
"BriefDescription": "Counts the number of far branch instructions retired.(Precise Event)",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{

4
tools/perf/pmu-events/arch/x86/broadwellde/cache.json
View file

@ -439,7 +439,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-split load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -451,7 +451,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-split store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"SampleAfterValue": "100003",
"L1_Hit_Indication": "1",
"CounterHTOff": "0,1,2,3"

6
tools/perf/pmu-events/arch/x86/broadwellde/pipeline.json
View file

@ -1,6 +1,5 @@
[
{
"EventCode": "0x00",
"UMask": "0x1",
"BriefDescription": "Instructions retired from execution.",
"Counter": "Fixed counter 0",
@ -10,7 +9,6 @@
"CounterHTOff": "Fixed counter 0"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when the thread is not in halt state",
"Counter": "Fixed counter 1",
@ -20,7 +18,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
"Counter": "Fixed counter 1",
@ -30,7 +27,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x3",
"BriefDescription": "Reference cycles when the core is not in halt state.",
"Counter": "Fixed counter 2",
@ -322,7 +318,7 @@
"BriefDescription": "Stalls caused by changing prefix length of the instruction.",
"Counter": "0,1,2,3",
"EventName": "ILD_STALL.LCP",
"PublicDescription": "This event counts stalls occurred due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
"PublicDescription": "This event counts stalls occured due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},

278
tools/perf/pmu-events/arch/x86/broadwellx/bdx-metrics.json
View file

@ -1,164 +1,370 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , IDQ.MITE_UOPS / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 16 * ( ICACHE.HIT + ICACHE.MISSES ) / 4.0 ) )",
"MetricGroup": "Frontend",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE.IFDATA_STALL - (( 14 * ITLB_MISSES.STLB_HIT + cpu@ITLB_MISSES.WALK_DURATION\\,cmask\\=1@ + 7* ITLB_MISSES.WALK_COMPLETED )) ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) * (12 * ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) ) * (4 * cycles) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "Branch_Misprediction_Cost"
},
{
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) * (12 * ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT + BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) ) * (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts_SMT",
"MetricName": "Branch_Misprediction_Cost_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_UOPS_RETIRED.L1_MISS + mem_load_uops_retired.hit_lfb )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( cpu@l1d_pend_miss.pending_cycles\\,any\\=1@ / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION + 7 * ( DTLB_STORE_MISSES.WALK_COMPLETED + DTLB_LOAD_MISSES.WALK_COMPLETED + ITLB_MISSES.WALK_COMPLETED ) ) / ( 2 * cycles )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION + 7*(DTLB_STORE_MISSES.WALK_COMPLETED+DTLB_LOAD_MISSES.WALK_COMPLETED+ITLB_MISSES.WALK_COMPLETED) ) / (2*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles))",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION + 7 * ( DTLB_STORE_MISSES.WALK_COMPLETED + DTLB_LOAD_MISSES.WALK_COMPLETED + ITLB_MISSES.WALK_COMPLETED ) ) / ( 2 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )) )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * L2_RQSTS.MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * ( L2_RQSTS.REFERENCES - L2_RQSTS.MISS ) / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2* FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4*( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8* FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "( 64 * ( uncore_imc@cas_count_read@ + uncore_imc@cas_count_write@ ) / 1000000000 ) / duration_time",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "1000000000 * ( cbox@event\\=0x36\\,umask\\=0x3\\,filter_opc\\=0x182@ / cbox@event\\=0x35\\,umask\\=0x3\\,filter_opc\\=0x182@ ) / ( cbox_0@event\\=0x0@ / duration_time )",
"BriefDescription": "Average latency of data read request to external memory (in nanoseconds). Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_Lat",
"MetricName": "DRAM_Read_Latency"
},
{
"MetricExpr": "cbox@event\\=0x36\\,umask\\=0x3\\,filter_opc\\=0x182@ / cbox@event\\=0x36\\,umask\\=0x3\\,filter_opc\\=0x182\\,thresh\\=1@",
"BriefDescription": "Average number of parallel data read requests to external memory. Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_Parallel_Reads"
},
{
"MetricExpr": "cbox_0@event\\=0x0@",
"BriefDescription": "Socket actual clocks when any core is active on that socket",
"MetricGroup": "",
"MetricName": "Socket_CLKS"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

161
tools/perf/pmu-events/arch/x86/broadwellx/cache.json
View file

@ -57,17 +57,17 @@
},
{
"EventCode": "0x24",
"UMask": "0x41",
"UMask": "0xc1",
"BriefDescription": "Demand Data Read requests that hit L2 cache",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
"PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
"PublicDescription": "Counts the number of demand Data Read requests, initiated by load instructions, that hit L2 cache.",
"SampleAfterValue": "200003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"UMask": "0x42",
"UMask": "0xc2",
"BriefDescription": "RFO requests that hit L2 cache.",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.RFO_HIT",
@ -76,7 +76,7 @@
},
{
"EventCode": "0x24",
"UMask": "0x44",
"UMask": "0xc4",
"BriefDescription": "L2 cache hits when fetching instructions, code reads.",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.CODE_RD_HIT",
@ -85,7 +85,7 @@
},
{
"EventCode": "0x24",
"UMask": "0x50",
"UMask": "0xd0",
"BriefDescription": "L2 prefetch requests that hit L2 cache",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.L2_PF_HIT",
@ -396,24 +396,24 @@
{
"EventCode": "0xD0",
"UMask": "0x11",
"BriefDescription": "Retired load uops that miss the STLB. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops that miss the STLB.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.STLB_MISS_LOADS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
"PublicDescription": "This event counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD0",
"UMask": "0x12",
"BriefDescription": "Retired store uops that miss the STLB. (Precise Event - PEBS)",
"BriefDescription": "Retired store uops that miss the STLB.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.STLB_MISS_STORES",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
"PublicDescription": "This event counts store uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
"SampleAfterValue": "100003",
"L1_Hit_Indication": "1",
"CounterHTOff": "0,1,2,3"
@ -421,37 +421,37 @@
{
"EventCode": "0xD0",
"UMask": "0x21",
"BriefDescription": "Retired load uops with locked access. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops with locked access.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
"Errata": "BDM35",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with locked access retired to the architected path.",
"PublicDescription": "This event counts load uops with locked access retired to the architected path.",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD0",
"UMask": "0x41",
"BriefDescription": "Retired load uops that split across a cacheline boundary.(Precise Event - PEBS)",
"BriefDescription": "Retired load uops that split across a cacheline boundary.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-split load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"PublicDescription": "This event counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD0",
"UMask": "0x42",
"BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
"BriefDescription": "Retired store uops that split across a cacheline boundary.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-split store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"PublicDescription": "This event counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"SampleAfterValue": "100003",
"L1_Hit_Indication": "1",
"CounterHTOff": "0,1,2,3"
@ -459,24 +459,24 @@
{
"EventCode": "0xD0",
"UMask": "0x81",
"BriefDescription": "All retired load uops. (Precise Event - PEBS)",
"BriefDescription": "All retired load uops.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
"PublicDescription": "This event counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD0",
"UMask": "0x82",
"BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
"BriefDescription": "All retired store uops.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.ALL_STORES",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
"PublicDescription": "This event counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
"SampleAfterValue": "2000003",
"L1_Hit_Indication": "1",
"CounterHTOff": "0,1,2,3"
@ -484,69 +484,69 @@
{
"EventCode": "0xD1",
"UMask": "0x1",
"BriefDescription": "Retired load uops with L1 cache hits as data sources. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops with L1 cache hits as data sources.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data source were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
"PublicDescription": "This event counts retired load uops which data sources were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD1",
"UMask": "0x2",
"BriefDescription": "Retired load uops with L2 cache hits as data sources. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops with L2 cache hits as data sources.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
"Errata": "BDM35",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
"PublicDescription": "This event counts retired load uops which data sources were hits in the mid-level (L2) cache.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD1",
"UMask": "0x4",
"BriefDescription": "Hit in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops which data sources were data hits in L3 without snoops required.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L3_HIT",
"Errata": "BDM100",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
"PublicDescription": "This event counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
"SampleAfterValue": "50021",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD1",
"UMask": "0x8",
"BriefDescription": "Retired load uops misses in L1 cache as data sources. Uses PEBS.",
"BriefDescription": "Retired load uops misses in L1 cache as data sources.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
"PublicDescription": "This event counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD1",
"UMask": "0x10",
"BriefDescription": "Retired load uops with L2 cache misses as data sources. Uses PEBS.",
"BriefDescription": "Miss in mid-level (L2) cache. Excludes Unknown data-source.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
"PublicDescription": "This event counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
"SampleAfterValue": "50021",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD1",
"UMask": "0x20",
"BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS).",
"BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -558,83 +558,84 @@
{
"EventCode": "0xD1",
"UMask": "0x40",
"BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.HIT_LFB",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
"PublicDescription": "This event counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD2",
"UMask": "0x1",
"BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS",
"Errata": "BDM100",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
"PublicDescription": "This event counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
"SampleAfterValue": "20011",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD2",
"UMask": "0x2",
"BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT",
"Errata": "BDM100",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
"PublicDescription": "This event counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
"SampleAfterValue": "20011",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD2",
"UMask": "0x4",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared L3. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared L3.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM",
"Errata": "BDM100",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
"PublicDescription": "This event counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
"SampleAfterValue": "20011",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD2",
"UMask": "0x8",
"BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required. (Precise Event - PEBS)",
"BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_NONE",
"Errata": "BDM100",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
"PublicDescription": "This event counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD3",
"UMask": "0x1",
"BriefDescription": "Data from local DRAM either Snoop not needed or Snoop Miss (RspI)",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM",
"Errata": "BDE70, BDM100",
"PublicDescription": "This event counts retired load uops where the data came from local DRAM. This does not include hardware prefetches. This is a precise event.",
"PublicDescription": "Retired load uop whose Data Source was: local DRAM either Snoop not needed or Snoop Miss (RspI).",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD3",
"UMask": "0x4",
"BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI) (Precise Event)",
"BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI)",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -646,7 +647,7 @@
{
"EventCode": "0xD3",
"UMask": "0x10",
"BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM (Precise Event)",
"BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -658,7 +659,7 @@
{
"EventCode": "0xD3",
"UMask": "0x20",
"BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache (Precise Event)",
"BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -810,12 +811,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all requests that hit in the L3",
"MSRValue": "0x3f803c8fff",
"BriefDescription": "Counts all requests hit in the L3",
"MSRValue": "0x3F803C8FFF",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all requests that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all requests hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -823,12 +824,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c07f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C07F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -836,12 +837,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c07f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C07F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -849,12 +850,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0244",
"BriefDescription": "Counts all demand & prefetch code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0244",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -862,12 +863,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0122",
"BriefDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -875,12 +876,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0122",
"BriefDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -888,12 +889,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0091",
"BriefDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -901,12 +902,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0091",
"BriefDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -914,12 +915,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3",
"MSRValue": "0x3f803c0200",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads hit in the L3",
"MSRValue": "0x3F803C0200",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_CODE_RD.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -927,12 +928,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3",
"MSRValue": "0x3f803c0100",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs hit in the L3",
"MSRValue": "0x3F803C0100",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_RFO.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -940,12 +941,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0002",
"BriefDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -953,12 +954,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3",
"MSRValue": "0x3f803c0002",
"BriefDescription": "Counts all demand data writes (RFOs) hit in the L3",
"MSRValue": "0x3F803C0002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
}

16
tools/perf/pmu-events/arch/x86/broadwellx/floating-point.json
View file

@ -42,7 +42,7 @@
{
"EventCode": "0xC7",
"UMask": "0x3",
"BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. (RSQRT for single precision?)",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.SCALAR",
"SampleAfterValue": "2000003",
@ -51,7 +51,7 @@
{
"EventCode": "0xC7",
"UMask": "0x4",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired. Each count represents 2 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired. Each count represents 2 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE",
"SampleAfterValue": "2000003",
@ -60,7 +60,7 @@
{
"EventCode": "0xC7",
"UMask": "0x8",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE",
"SampleAfterValue": "2000003",
@ -69,7 +69,7 @@
{
"EventCode": "0xC7",
"UMask": "0x10",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE",
"SampleAfterValue": "2000003",
@ -78,7 +78,7 @@
{
"EventCode": "0xC7",
"UMask": "0x15",
"BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
"BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
"SampleAfterValue": "2000006",
@ -87,7 +87,7 @@
{
"EventCode": "0xc7",
"UMask": "0x20",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
"SampleAfterValue": "2000003",
@ -96,7 +96,7 @@
{
"EventCode": "0xC7",
"UMask": "0x2a",
"BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
"BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.SINGLE",
"SampleAfterValue": "2000005",
@ -105,7 +105,7 @@
{
"EventCode": "0xC7",
"UMask": "0x3c",
"BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
"BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. (RSQRT for single-precision?)",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.PACKED",
"SampleAfterValue": "2000004",

148
tools/perf/pmu-events/arch/x86/broadwellx/memory.json
View file

@ -170,11 +170,11 @@
{
"EventCode": "0xc8",
"UMask": "0x4",
"BriefDescription": "Number of times HLE abort was triggered (PEBS)",
"BriefDescription": "Number of times HLE abort was triggered",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "HLE_RETIRED.ABORTED",
"PublicDescription": "Number of times HLE abort was triggered (PEBS).",
"PublicDescription": "Number of times HLE abort was triggered.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -251,11 +251,11 @@
{
"EventCode": "0xc9",
"UMask": "0x4",
"BriefDescription": "Number of times RTM abort was triggered (PEBS)",
"BriefDescription": "Number of times RTM abort was triggered",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "RTM_RETIRED.ABORTED",
"PublicDescription": "Number of times RTM abort was triggered (PEBS).",
"PublicDescription": "Number of times RTM abort was triggered .",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3"
},
@ -312,14 +312,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 4",
"BriefDescription": "Randomly selected loads with latency value being above 4",
"PEBS": "2",
"MSRValue": "0x4",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_4",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above four.",
"PublicDescription": "Counts randomly selected loads with latency value being above four.",
"TakenAlone": "1",
"SampleAfterValue": "100003",
"CounterHTOff": "3"
@ -327,14 +327,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 8",
"BriefDescription": "Randomly selected loads with latency value being above 8",
"PEBS": "2",
"MSRValue": "0x8",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_8",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above eight.",
"PublicDescription": "Counts randomly selected loads with latency value being above eight.",
"TakenAlone": "1",
"SampleAfterValue": "50021",
"CounterHTOff": "3"
@ -342,14 +342,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 16",
"BriefDescription": "Randomly selected loads with latency value being above 16",
"PEBS": "2",
"MSRValue": "0x10",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_16",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above 16.",
"PublicDescription": "Counts randomly selected loads with latency value being above 16.",
"TakenAlone": "1",
"SampleAfterValue": "20011",
"CounterHTOff": "3"
@ -357,14 +357,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 32",
"BriefDescription": "Randomly selected loads with latency value being above 32",
"PEBS": "2",
"MSRValue": "0x20",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_32",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above 32.",
"PublicDescription": "Counts randomly selected loads with latency value being above 32.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "3"
@ -372,14 +372,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 64",
"BriefDescription": "Randomly selected loads with latency value being above 64",
"PEBS": "2",
"MSRValue": "0x40",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_64",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above 64.",
"PublicDescription": "Counts randomly selected loads with latency value being above 64.",
"TakenAlone": "1",
"SampleAfterValue": "2003",
"CounterHTOff": "3"
@ -387,14 +387,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 128",
"BriefDescription": "Randomly selected loads with latency value being above 128",
"PEBS": "2",
"MSRValue": "0x80",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_128",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above 128.",
"PublicDescription": "Counts randomly selected loads with latency value being above 128.",
"TakenAlone": "1",
"SampleAfterValue": "1009",
"CounterHTOff": "3"
@ -402,14 +402,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 256",
"BriefDescription": "Randomly selected loads with latency value being above 256",
"PEBS": "2",
"MSRValue": "0x100",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_256",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above 256.",
"PublicDescription": "Counts randomly selected loads with latency value being above 256.",
"TakenAlone": "1",
"SampleAfterValue": "503",
"CounterHTOff": "3"
@ -417,14 +417,14 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 512",
"BriefDescription": "Randomly selected loads with latency value being above 512",
"PEBS": "2",
"MSRValue": "0x200",
"Counter": "3",
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_512",
"MSRIndex": "0x3F6",
"Errata": "BDM100, BDM35",
"PublicDescription": "This event counts loads with latency value being above 512.",
"PublicDescription": "Counts randomly selected loads with latency value being above 512.",
"TakenAlone": "1",
"SampleAfterValue": "101",
"CounterHTOff": "3"
@ -433,12 +433,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all requests that miss in the L3",
"MSRValue": "0x3fbfc08fff",
"BriefDescription": "Counts all requests miss in the L3",
"MSRValue": "0x3FBFC08FFF",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all requests that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all requests miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -446,12 +446,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x087fc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x087FC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_HIT_FORWARD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and clean or shared data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and clean or shared data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -459,12 +459,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103fc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103FC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_HITM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the modified data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -472,12 +472,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063bc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063BC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from remote dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from remote dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -485,12 +485,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram",
"MSRValue": "0x06040007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from local dram",
"MSRValue": "0x06040007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -498,12 +498,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3",
"MSRValue": "0x3fbfc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss in the L3",
"MSRValue": "0x3FBFC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -511,12 +511,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch code reads miss the L3 and the data is returned from local dram",
"MSRValue": "0x0604000244",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -524,12 +524,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch code reads that miss in the L3",
"MSRValue": "0x3fbfc00244",
"BriefDescription": "Counts all demand & prefetch code reads miss in the L3",
"MSRValue": "0x3FBFC00244",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -537,12 +537,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch RFOs miss the L3 and the data is returned from local dram",
"MSRValue": "0x0604000122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -550,12 +550,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that miss in the L3",
"MSRValue": "0x3fbfc00122",
"BriefDescription": "Counts all demand & prefetch RFOs miss in the L3",
"MSRValue": "0x3FBFC00122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -563,12 +563,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x087fc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x087FC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_HIT_FORWARD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and clean or shared data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and clean or shared data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -576,12 +576,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103fc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103FC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_HITM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the modified data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -589,12 +589,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063bc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063BC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from remote dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from remote dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -602,12 +602,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from local dram",
"MSRValue": "0x0604000091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -615,12 +615,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss in the L3",
"MSRValue": "0x3fbfc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss in the L3",
"MSRValue": "0x3FBFC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -628,12 +628,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3",
"MSRValue": "0x3fbfc00200",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads miss in the L3",
"MSRValue": "0x3FBFC00200",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_CODE_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -641,12 +641,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3",
"MSRValue": "0x3fbfc00100",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs miss in the L3",
"MSRValue": "0x3FBFC00100",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -654,12 +654,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103fc00002",
"BriefDescription": "Counts all demand data writes (RFOs) miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103FC00002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.REMOTE_HITM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss the L3 and the modified data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -667,12 +667,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that miss in the L3",
"MSRValue": "0x3fbfc00002",
"BriefDescription": "Counts all demand data writes (RFOs) miss in the L3",
"MSRValue": "0x3FBFC00002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
}

50
tools/perf/pmu-events/arch/x86/broadwellx/pipeline.json
View file

@ -1,6 +1,5 @@
[
{
"EventCode": "0x00",
"UMask": "0x1",
"BriefDescription": "Instructions retired from execution.",
"Counter": "Fixed counter 0",
@ -10,7 +9,6 @@
"CounterHTOff": "Fixed counter 0"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when the thread is not in halt state",
"Counter": "Fixed counter 1",
@ -20,7 +18,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
"Counter": "Fixed counter 1",
@ -30,7 +27,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x3",
"BriefDescription": "Reference cycles when the core is not in halt state.",
"Counter": "Fixed counter 2",
@ -322,7 +318,7 @@
"BriefDescription": "Stalls caused by changing prefix length of the instruction.",
"Counter": "0,1,2,3",
"EventName": "ILD_STALL.LCP",
"PublicDescription": "This event counts stalls occurred due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
"PublicDescription": "This event counts stalls occured due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -786,12 +782,12 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xA2",
"EventCode": "0xa2",
"UMask": "0x1",
"BriefDescription": "Resource-related stall cycles",
"Counter": "0,1,2,3",
"EventName": "RESOURCE_STALLS.ANY",
"PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
"PublicDescription": "This event counts resource-related stall cycles.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1168,12 +1164,12 @@
{
"EventCode": "0xC2",
"UMask": "0x1",
"BriefDescription": "Actually retired uops. (Precise Event - PEBS)",
"BriefDescription": "Actually retired uops.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.ALL",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
"PublicDescription": "This event counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1204,11 +1200,11 @@
{
"EventCode": "0xC2",
"UMask": "0x2",
"BriefDescription": "Retirement slots used. (Precise Event - PEBS)",
"BriefDescription": "Retirement slots used.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.RETIRE_SLOTS",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of retirement slots used.",
"PublicDescription": "This event counts the number of retirement slots used.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1266,33 +1262,33 @@
{
"EventCode": "0xC4",
"UMask": "0x1",
"BriefDescription": "Conditional branch instructions retired. (Precise Event - PEBS)",
"BriefDescription": "Conditional branch instructions retired.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.CONDITIONAL",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts conditional branch instructions retired.",
"PublicDescription": "This event counts conditional branch instructions retired.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC4",
"UMask": "0x2",
"BriefDescription": "Direct and indirect near call instructions retired. (Precise Event - PEBS)",
"BriefDescription": "Direct and indirect near call instructions retired.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NEAR_CALL",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect near call instructions retired.",
"PublicDescription": "This event counts both direct and indirect near call instructions retired.",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC4",
"UMask": "0x2",
"BriefDescription": "Direct and indirect macro near call instructions retired (captured in ring 3). (Precise Event - PEBS)",
"BriefDescription": "Direct and indirect macro near call instructions retired (captured in ring 3).",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NEAR_CALL_R3",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect macro near call instructions retired (captured in ring 3).",
"PublicDescription": "This event counts both direct and indirect macro near call instructions retired (captured in ring 3).",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1311,11 +1307,11 @@
{
"EventCode": "0xC4",
"UMask": "0x8",
"BriefDescription": "Return instructions retired. (Precise Event - PEBS)",
"BriefDescription": "Return instructions retired.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NEAR_RETURN",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts return instructions retired.",
"PublicDescription": "This event counts return instructions retired.",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1332,11 +1328,11 @@
{
"EventCode": "0xC4",
"UMask": "0x20",
"BriefDescription": "Taken branch instructions retired. (Precise Event - PEBS)",
"BriefDescription": "Taken branch instructions retired.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NEAR_TAKEN",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts taken branch instructions retired.",
"PublicDescription": "This event counts taken branch instructions retired.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1364,11 +1360,11 @@
{
"EventCode": "0xC5",
"UMask": "0x1",
"BriefDescription": "Mispredicted conditional branch instructions retired. (Precise Event - PEBS)",
"BriefDescription": "Mispredicted conditional branch instructions retired.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_MISP_RETIRED.CONDITIONAL",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted conditional branch instructions retired.",
"PublicDescription": "This event counts mispredicted conditional branch instructions retired.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1386,22 +1382,22 @@
{
"EventCode": "0xC5",
"UMask": "0x8",
"BriefDescription": "This event counts the number of mispredicted ret instructions retired.(Precise Event)",
"BriefDescription": "This event counts the number of mispredicted ret instructions retired. Non PEBS",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_MISP_RETIRED.RET",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted return instructions retired.",
"PublicDescription": "This event counts mispredicted return instructions retired.",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC5",
"UMask": "0x20",
"BriefDescription": "number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
"BriefDescription": "number of near branch instructions retired that were mispredicted and taken.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
"PublicDescription": "Number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
"PublicDescription": "Number of near branch instructions retired that were mispredicted and taken.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},

304
tools/perf/pmu-events/arch/x86/cascadelakex/clx-metrics.json
View file

@ -1,164 +1,394 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ((UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 64 * ( ICACHE_64B.IFTAG_HIT + ICACHE_64B.IFTAG_MISS ) / 4.1) )",
"MetricGroup": "Frontend",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 64 * ( ICACHE_64B.IFTAG_HIT + ICACHE_64B.IFTAG_MISS ) / 4.1 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ))",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_INST_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_INST_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * ( FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE ) + 16 * FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * ( FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE ) + 16 * FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( UOPS_EXECUTED.CORE_CYCLES_GE_1 / 2 ) if #SMT_on else UOPS_EXECUTED.CORE_CYCLES_GE_1)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( UOPS_EXECUTED.CORE_CYCLES_GE_1 / 2) if #SMT_on else UOPS_EXECUTED.CORE_CYCLES_GE_1)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE_16B.IFDATA_STALL - ICACHE_64B.IFTAG_STALL ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) * (( INT_MISC.CLEAR_RESTEER_CYCLES + 9 * BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) ) * (4 * cycles) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "Branch_Misprediction_Cost"
},
{
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) * (( INT_MISC.CLEAR_RESTEER_CYCLES + 9 * BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) ) * (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts_SMT",
"MetricName": "Branch_Misprediction_Cost_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_RETIRED.L1_MISS + MEM_LOAD_RETIRED.FB_HIT )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( L1D_PEND_MISS.PENDING_CYCLES_ANY / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * cycles )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles) )",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )) )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "64 * OFFCORE_REQUESTS.ALL_REQUESTS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Access_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * L2_RQSTS.MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * ( L2_RQSTS.REFERENCES - L2_RQSTS.MISS ) / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * ( FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE ) + 16 * FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2* FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4*( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8* FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "( 64 * ( uncore_imc@cas_count_read@ + uncore_imc@cas_count_write@ ) / 1000000000 ) / duration_time",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "1000000000 * ( cha@event\\=0x36\\\\\\,umask\\=0x21@ / cha@event\\=0x35\\\\\\,umask\\=0x21@ ) / ( cha_0@event\\=0x0@ / duration_time )",
"BriefDescription": "Average latency of data read request to external memory (in nanoseconds). Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_Lat",
"MetricName": "DRAM_Read_Latency"
},
{
"MetricExpr": "cha@event\\=0x36\\\\\\,umask\\=0x21@ / cha@event\\=0x36\\\\\\,umask\\=0x21\\\\\\,thresh\\=1@",
"BriefDescription": "Average number of parallel data read requests to external memory. Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_Parallel_Reads"
},
{
"MetricExpr": "( 1000000000 * ( imc@event\\=0xe0\\\\\\,umask\\=0x1@ / imc@event\\=0xe3@ ) / imc_0@event\\=0x0@ ) if 1 if 1 == 1 else 0 else 0",
"BriefDescription": "Average latency of data read request to external 3D X-Point memory [in nanoseconds]. Accounts for demand loads and L1/L2 data-read prefetches",
"MetricGroup": "Memory_Lat",
"MetricName": "MEM_PMM_Read_Latency"
},
{
"MetricExpr": "( ( 64 * imc@event\\=0xe3@ / 1000000000 ) / duration_time ) if 1 if 1 == 1 else 0 else 0",
"BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "PMM_Read_BW"
},
{
"MetricExpr": "( ( 64 * imc@event\\=0xe7@ / 1000000000 ) / duration_time ) if 1 if 1 == 1 else 0 else 0",
"BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "PMM_Write_BW"
},
{
"MetricExpr": "cha_0@event\\=0x0@",
"BriefDescription": "Socket actual clocks when any core is active on that socket",
"MetricGroup": "",
"MetricName": "Socket_CLKS"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

980
tools/perf/pmu-events/arch/x86/goldmont/cache.json
View file

File diff suppressed because it is too large Load diff

260
tools/perf/pmu-events/arch/x86/goldmont/memory.json
View file

@ -30,265 +30,5 @@
"EventName": "MACHINE_CLEARS.MEMORY_ORDERING",
"SampleAfterValue": "200003",
"BriefDescription": "Machine clears due to memory ordering issue"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x20000032b7 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000022 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts data reads (demand & prefetch) that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000003091",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data reads (demand & prefetch) that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000003010 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_PF_DATA_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data reads generated by L1 or L2 prefetchers that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts requests to the uncore subsystem that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000008000 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts requests to the uncore subsystem that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts any data writes to uncacheable write combining (USWC) memory region that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000004800 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.STREAMING_STORES.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any data writes to uncacheable write combining (USWC) memory region that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts partial cache line data writes to uncacheable write combining (USWC) memory region that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000004000 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_STREAMING_STORES.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts partial cache line data writes to uncacheable write combining (USWC) memory region that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts data cache line reads generated by hardware L1 data cache prefetcher that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000002000 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L1_DATA_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data cache line reads generated by hardware L1 data cache prefetcher that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts data cache lines requests by software prefetch instructions that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000001000 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.SW_PREFETCH.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data cache lines requests by software prefetch instructions that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts full cache line data writes to uncacheable write combining (USWC) memory region and full cache-line non-temporal writes that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000800 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.FULL_STREAMING_STORES.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts full cache line data writes to uncacheable write combining (USWC) memory region and full cache-line non-temporal writes that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts bus lock and split lock requests that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000400 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts bus lock and split lock requests that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts code reads in uncacheable (UC) memory region that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000200 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts code reads in uncacheable (UC) memory region that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of demand write requests (RFO) generated by a write to partial data cache line, including the writes to uncacheable (UC) and write through (WT), and write protected (WP) types of memory that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000100 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_WRITES.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts the number of demand write requests (RFO) generated by a write to partial data cache line, including the writes to uncacheable (UC) and write through (WT), and write protected (WP) types of memory that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts demand data partial reads, including data in uncacheable (UC) or uncacheable write combining (USWC) memory types that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000080 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand data partial reads, including data in uncacheable (UC) or uncacheable write combining (USWC) memory types that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts reads for ownership (RFO) requests generated by L2 prefetcher that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000020 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts reads for ownership (RFO) requests generated by L2 prefetcher that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts data cacheline reads generated by hardware L2 cache prefetcher that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000010 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_DATA_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data cacheline reads generated by hardware L2 cache prefetcher that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of writeback transactions caused by L1 or L2 cache evictions that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000008 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.COREWB.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6",
"SampleAfterValue": "100007",
"BriefDescription": "Counts the number of writeback transactions caused by L1 or L2 cache evictions that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts demand instruction cacheline and I-side prefetch requests that miss the instruction cache that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000004 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand instruction cacheline and I-side prefetch requests that miss the instruction cache that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts demand reads for ownership (RFO) requests generated by a write to full data cache line that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000002 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand reads for ownership (RFO) requests generated by a write to full data cache line that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts demand cacheable data reads of full cache lines that miss the L2 cache and targets non-DRAM system address. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
"EventCode": "0xB7",
"MSRValue": "0x2000000001 ",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.L2_MISS.NON_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand cacheable data reads of full cache lines that miss the L2 cache and targets non-DRAM system address.",
"Offcore": "1"
}
]

5
tools/perf/pmu-events/arch/x86/goldmont/pipeline.json
View file

@ -1,7 +1,6 @@
[
{
"PublicDescription": "Counts the number of instructions that retire execution. For instructions that consist of multiple uops, this event counts the retirement of the last uop of the instruction. The counter continues counting during hardware interrupts, traps, and inside interrupt handlers. This event uses fixed counter 0. You cannot collect a PEBs record for this event.",
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -10,7 +9,6 @@
},
{
"PublicDescription": "Counts the number of core cycles while the core is not in a halt state. The core enters the halt state when it is running the HLT instruction. In mobile systems the core frequency may change from time to time. For this reason this event may have a changing ratio with regards to time. This event uses fixed counter 1. You cannot collect a PEBs record for this event.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.CORE",
@ -19,7 +17,6 @@
},
{
"PublicDescription": "Counts the number of reference cycles that the core is not in a halt state. The core enters the halt state when it is running the HLT instruction. In mobile systems the core frequency may change from time. This event is not affected by core frequency changes but counts as if the core is running at the maximum frequency all the time. This event uses fixed counter 2. You cannot collect a PEBs record for this event.",
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",
@ -188,7 +185,7 @@
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of times that the processor detects that a program is writing to a code section and has to perform a machine clear because of that modification. Self-modifying code (SMC) causes a severe penalty in all Intel architecture processors.",
"PublicDescription": "Counts the number of times that the processor detects that a program is writing to a code section and has to perform a machine clear because of that modification. Self-modifying code (SMC) causes a severe penalty in all Intel\u00ae architecture processors.",
"EventCode": "0xC3",
"Counter": "0,1,2,3",
"UMask": "0x1",

9
tools/perf/pmu-events/arch/x86/goldmont/virtual-memory.json
View file

@ -48,7 +48,8 @@
"UMask": "0x11",
"EventName": "MEM_UOPS_RETIRED.DTLB_MISS_LOADS",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that missed the DTLB (Precise event capable)"
"BriefDescription": "Load uops retired that missed the DTLB (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -59,7 +60,8 @@
"UMask": "0x12",
"EventName": "MEM_UOPS_RETIRED.DTLB_MISS_STORES",
"SampleAfterValue": "200003",
"BriefDescription": "Store uops retired that missed the DTLB (Precise event capable)"
"BriefDescription": "Store uops retired that missed the DTLB (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -70,6 +72,7 @@
"UMask": "0x13",
"EventName": "MEM_UOPS_RETIRED.DTLB_MISS",
"SampleAfterValue": "200003",
"BriefDescription": "Memory uops retired that missed the DTLB (Precise event capable)"
"BriefDescription": "Memory uops retired that missed the DTLB (Precise event capable)",
"Data_LA": "1"
}
]

74
tools/perf/pmu-events/arch/x86/goldmontplus/cache.json
View file

@ -92,7 +92,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
"SampleAfterValue": "200003",
"BriefDescription": "Locked load uops retired (Precise event capable)"
"BriefDescription": "Locked load uops retired (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -104,7 +105,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that split a cache-line (Precise event capable)"
"BriefDescription": "Load uops retired that split a cache-line (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -116,7 +118,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
"SampleAfterValue": "200003",
"BriefDescription": "Stores uops retired that split a cache-line (Precise event capable)"
"BriefDescription": "Stores uops retired that split a cache-line (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -128,7 +131,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT",
"SampleAfterValue": "200003",
"BriefDescription": "Memory uops retired that split a cache-line (Precise event capable)"
"BriefDescription": "Memory uops retired that split a cache-line (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -140,7 +144,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired (Precise event capable)"
"BriefDescription": "Load uops retired (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -152,7 +157,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.ALL_STORES",
"SampleAfterValue": "200003",
"BriefDescription": "Store uops retired (Precise event capable)"
"BriefDescription": "Store uops retired (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -164,7 +170,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.ALL",
"SampleAfterValue": "200003",
"BriefDescription": "Memory uops retired (Precise event capable)"
"BriefDescription": "Memory uops retired (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -176,7 +183,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that hit L1 data cache (Precise event capable)"
"BriefDescription": "Load uops retired that hit L1 data cache (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -188,7 +196,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that hit L2 (Precise event capable)"
"BriefDescription": "Load uops retired that hit L2 (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -200,7 +209,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that missed L1 data cache (Precise event capable)"
"BriefDescription": "Load uops retired that missed L1 data cache (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -212,7 +222,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that missed L2 (Precise event capable)"
"BriefDescription": "Load uops retired that missed L2 (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -224,7 +235,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.HITM",
"SampleAfterValue": "200003",
"BriefDescription": "Memory uop retired where cross core or cross module HITM occurred (Precise event capable)"
"BriefDescription": "Memory uop retired where cross core or cross module HITM occurred (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -236,7 +248,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.WCB_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "Loads retired that hit WCB (Precise event capable)"
"BriefDescription": "Loads retired that hit WCB (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -248,7 +261,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.DRAM_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "Loads retired that came from DRAM (Precise event capable)"
"BriefDescription": "Loads retired that came from DRAM (Precise event capable)",
"Data_LA": "1"
},
{
"CollectPEBSRecord": "1",
@ -292,7 +306,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand cacheable data reads of full cache lines true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts demand cacheable data reads of full cache lines true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -367,7 +381,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand reads for ownership (RFO) requests generated by a write to full data cache line true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts demand reads for ownership (RFO) requests generated by a write to full data cache line true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -442,7 +456,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand instruction cacheline and I-side prefetch requests that miss the instruction cache true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts demand instruction cacheline and I-side prefetch requests that miss the instruction cache true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -517,7 +531,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts the number of writeback transactions caused by L1 or L2 cache evictions true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts the number of writeback transactions caused by L1 or L2 cache evictions true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -592,7 +606,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data cacheline reads generated by hardware L2 cache prefetcher true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts data cacheline reads generated by hardware L2 cache prefetcher true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -667,7 +681,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts reads for ownership (RFO) requests generated by L2 prefetcher true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts reads for ownership (RFO) requests generated by L2 prefetcher true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -742,7 +756,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts bus lock and split lock requests true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts bus lock and split lock requests true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -817,7 +831,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts full cache line data writes to uncacheable write combining (USWC) memory region and full cache-line non-temporal writes true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts full cache line data writes to uncacheable write combining (USWC) memory region and full cache-line non-temporal writes true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -892,7 +906,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data cache lines requests by software prefetch instructions true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts data cache lines requests by software prefetch instructions true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -967,7 +981,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data cache line reads generated by hardware L1 data cache prefetcher true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts data cache line reads generated by hardware L1 data cache prefetcher true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -1042,7 +1056,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any data writes to uncacheable write combining (USWC) memory region true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts any data writes to uncacheable write combining (USWC) memory region true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -1117,7 +1131,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts requests to the uncore subsystem true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts requests to the uncore subsystem true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -1192,7 +1206,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data reads generated by L1 or L2 prefetchers true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts data reads generated by L1 or L2 prefetchers true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -1267,7 +1281,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data reads (demand & prefetch) true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts data reads (demand & prefetch) true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -1342,7 +1356,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{
@ -1417,7 +1431,7 @@
"PDIR_COUNTER": "na",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) true miss for the L2 cache with a snoop miss in the other processor module. ",
"BriefDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) true miss for the L2 cache with a snoop miss in the other processor module.",
"Offcore": "1"
},
{

5
tools/perf/pmu-events/arch/x86/goldmontplus/pipeline.json
View file

@ -3,7 +3,6 @@
"PEBS": "2",
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of instructions that retire execution. For instructions that consist of multiple uops, this event counts the retirement of the last uop of the instruction. The counter continues counting during hardware interrupts, traps, and inside interrupt handlers. This event uses fixed counter 0. You cannot collect a PEBs record for this event.",
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"PEBScounters": "32",
@ -15,7 +14,6 @@
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of core cycles while the core is not in a halt state. The core enters the halt state when it is running the HLT instruction. In mobile systems the core frequency may change from time to time. For this reason this event may have a changing ratio with regards to time. This event uses fixed counter 1. You cannot collect a PEBs record for this event.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"PEBScounters": "33",
@ -27,7 +25,6 @@
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of reference cycles that the core is not in a halt state. The core enters the halt state when it is running the HLT instruction. In mobile systems the core frequency may change from time. This event is not affected by core frequency changes but counts as if the core is running at the maximum frequency all the time. This event uses fixed counter 2. You cannot collect a PEBs record for this event.",
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"PEBScounters": "34",
@ -231,7 +228,7 @@
},
{
"CollectPEBSRecord": "1",
"PublicDescription": "Counts the number of times that the processor detects that a program is writing to a code section and has to perform a machine clear because of that modification. Self-modifying code (SMC) causes a severe penalty in all Intel architecture processors.",
"PublicDescription": "Counts the number of times that the processor detects that a program is writing to a code section and has to perform a machine clear because of that modification. Self-modifying code (SMC) causes a severe penalty in all Intel\u00ae architecture processors.",
"EventCode": "0xC3",
"Counter": "0,1,2,3",
"UMask": "0x1",

9
tools/perf/pmu-events/arch/x86/goldmontplus/virtual-memory.json
View file

@ -189,7 +189,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.DTLB_MISS_LOADS",
"SampleAfterValue": "200003",
"BriefDescription": "Load uops retired that missed the DTLB (Precise event capable)"
"BriefDescription": "Load uops retired that missed the DTLB (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -201,7 +202,8 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.DTLB_MISS_STORES",
"SampleAfterValue": "200003",
"BriefDescription": "Store uops retired that missed the DTLB (Precise event capable)"
"BriefDescription": "Store uops retired that missed the DTLB (Precise event capable)",
"Data_LA": "1"
},
{
"PEBS": "2",
@ -213,6 +215,7 @@
"PEBScounters": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.DTLB_MISS",
"SampleAfterValue": "200003",
"BriefDescription": "Memory uops retired that missed the DTLB (Precise event capable)"
"BriefDescription": "Memory uops retired that missed the DTLB (Precise event capable)",
"Data_LA": "1"
}
]

175
tools/perf/pmu-events/arch/x86/haswell/cache.json
View file

@ -63,10 +63,10 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Demand data read requests that hit L2 cache.",
"PublicDescription": "Counts the number of demand Data Read requests, initiated by load instructions, that hit L2 cache",
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x41",
"UMask": "0xc1",
"Errata": "HSD78",
"EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
"SampleAfterValue": "200003",
@ -77,7 +77,7 @@
"PublicDescription": "Counts the number of store RFO requests that hit the L2 cache.",
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x42",
"UMask": "0xc2",
"EventName": "L2_RQSTS.RFO_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "RFO requests that hit L2 cache",
@ -87,7 +87,7 @@
"PublicDescription": "Number of instruction fetches that hit the L2 cache.",
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x44",
"UMask": "0xc4",
"EventName": "L2_RQSTS.CODE_RD_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "L2 cache hits when fetching instructions, code reads.",
@ -97,7 +97,7 @@
"PublicDescription": "Counts all L2 HW prefetcher requests that hit L2.",
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x50",
"UMask": "0xd0",
"EventName": "L2_RQSTS.L2_PF_HIT",
"SampleAfterValue": "200003",
"BriefDescription": "L2 prefetch requests that hit L2 cache",
@ -610,7 +610,7 @@
"Errata": "HSD29, HSD25, HSM26, HSM30",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache. ",
"BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache.",
"CounterHTOff": "0,1,2,3",
"Data_LA": "1"
},
@ -623,7 +623,7 @@
"Errata": "HSD29, HSD25, HSM26, HSM30",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared L3. ",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared L3.",
"CounterHTOff": "0,1,2,3",
"Data_LA": "1"
},
@ -792,7 +792,6 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "",
"EventCode": "0xf4",
"Counter": "0,1,2,3",
"UMask": "0x10",
@ -802,262 +801,262 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts all requests that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all requests hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c8fff",
"MSRValue": "0x3F803C8FFF",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all requests that hit in the L3",
"BriefDescription": "Counts all requests hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c07f7",
"MSRValue": "0x10003C07F7",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_READS.L3_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"BriefDescription": "hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c07f7",
"MSRValue": "0x04003C07F7",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_READS.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c0244",
"MSRValue": "0x04003C0244",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "Counts all demand & prefetch code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c0122",
"MSRValue": "0x10003C0122",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.L3_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"BriefDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c0122",
"MSRValue": "0x04003C0122",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c0091",
"MSRValue": "0x10003C0091",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.L3_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"BriefDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c0091",
"MSRValue": "0x04003C0091",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c0200",
"MSRValue": "0x3F803C0200",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L3_CODE_RD.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c0100",
"MSRValue": "0x3F803C0100",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L3_RFO.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c0080",
"MSRValue": "0x3F803C0080",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L3_DATA_RD.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c0040",
"MSRValue": "0x3F803C0040",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c0020",
"MSRValue": "0x3F803C0020",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to L2) data reads hit in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3f803c0010",
"MSRValue": "0x3F803C0010",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_DATA_RD.L3_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads hit in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand code reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c0004",
"MSRValue": "0x10003C0004",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.L3_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand code reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"BriefDescription": "Counts all demand code reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c0004",
"MSRValue": "0x04003C0004",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "Counts all demand code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c0002",
"MSRValue": "0x10003C0002",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.L3_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"BriefDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c0002",
"MSRValue": "0x04003C0002",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts demand data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c0001",
"MSRValue": "0x10003C0001",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.L3_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts demand data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"BriefDescription": "Counts demand data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts demand data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x04003c0001",
"MSRValue": "0x04003C0001",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.L3_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts demand data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"BriefDescription": "Counts demand data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
}

33
tools/perf/pmu-events/arch/x86/haswell/floating-point.json
View file

@ -1,22 +1,26 @@
[
{
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x8",
"Errata": "HSD56, HSM57",
"EventName": "OTHER_ASSISTS.AVX_TO_SSE",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.",
"BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x10",
"Errata": "HSD56, HSM57",
"EventName": "OTHER_ASSISTS.SSE_TO_AVX",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.",
"BriefDescription": "Number of transitions from legacy SSE to AVX-256 when penalty applicable",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -30,53 +34,58 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of X87 FP assists due to output values.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "FP_ASSIST.X87_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to output value.",
"BriefDescription": "output - Numeric Overflow, Numeric Underflow, Inexact Result",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of X87 FP assists due to input values.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "FP_ASSIST.X87_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to input value.",
"BriefDescription": "input - Invalid Operation, Denormal Operand, SNaN Operand",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of SIMD FP assists due to output values.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "FP_ASSIST.SIMD_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to Output values",
"BriefDescription": "SSE* FP micro-code assist when output value is invalid.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of SIMD FP assists due to input values.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "FP_ASSIST.SIMD_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to input values",
"BriefDescription": "Any input SSE* FP Assist",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Cycles with any input/output SSE* or FP assists.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x1e",
"EventName": "FP_ASSIST.ANY",
"SampleAfterValue": "100003",
"BriefDescription": "Cycles with any input/output SSE or FP assist",
"BriefDescription": "Counts any FP_ASSIST umask was incrementing",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
}

234
tools/perf/pmu-events/arch/x86/haswell/hsw-metrics.json
View file

@ -1,158 +1,322 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , IDQ.MITE_UOPS / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 16 * ( ICACHE.HIT + ICACHE.MISSES ) / 4.0 ) )",
"MetricGroup": "Frontend",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "( UOPS_EXECUTED.CORE / 2 / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@) ) if #SMT_on else UOPS_EXECUTED.CORE / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "( UOPS_EXECUTED.CORE / 2 / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@) ) if #SMT_on else UOPS_EXECUTED.CORE / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE.IFDATA_STALL - (( 14 * ITLB_MISSES.STLB_HIT + ITLB_MISSES.WALK_DURATION )) ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_UOPS_RETIRED.L1_MISS + mem_load_uops_retired.hit_lfb )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( cpu@l1d_pend_miss.pending_cycles\\,any\\=1@ / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / cycles",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "64 * ( arb@event\\=0x81\\,umask\\=0x1@ + arb@event\\=0x84\\,umask\\=0x1@ ) / 1000000 / duration_time / 1000",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

172
tools/perf/pmu-events/arch/x86/haswell/memory.json
View file

@ -298,7 +298,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_4",
"MSRIndex": "0x3F6",
"SampleAfterValue": "100003",
"BriefDescription": "Loads with latency value being above 4.",
"BriefDescription": "Randomly selected loads with latency value being above 4.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -312,7 +312,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_8",
"MSRIndex": "0x3F6",
"SampleAfterValue": "50021",
"BriefDescription": "Loads with latency value being above 8.",
"BriefDescription": "Randomly selected loads with latency value being above 8.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -326,7 +326,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_16",
"MSRIndex": "0x3F6",
"SampleAfterValue": "20011",
"BriefDescription": "Loads with latency value being above 16.",
"BriefDescription": "Randomly selected loads with latency value being above 16.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -340,7 +340,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_32",
"MSRIndex": "0x3F6",
"SampleAfterValue": "100003",
"BriefDescription": "Loads with latency value being above 32.",
"BriefDescription": "Randomly selected loads with latency value being above 32.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -354,7 +354,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_64",
"MSRIndex": "0x3F6",
"SampleAfterValue": "2003",
"BriefDescription": "Loads with latency value being above 64.",
"BriefDescription": "Randomly selected loads with latency value being above 64.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -368,7 +368,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_128",
"MSRIndex": "0x3F6",
"SampleAfterValue": "1009",
"BriefDescription": "Loads with latency value being above 128.",
"BriefDescription": "Randomly selected loads with latency value being above 128.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -382,7 +382,7 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_256",
"MSRIndex": "0x3F6",
"SampleAfterValue": "503",
"BriefDescription": "Loads with latency value being above 256.",
"BriefDescription": "Randomly selected loads with latency value being above 256.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
@ -396,280 +396,280 @@
"EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_512",
"MSRIndex": "0x3F6",
"SampleAfterValue": "101",
"BriefDescription": "Loads with latency value being above 512.",
"BriefDescription": "Randomly selected loads with latency value being above 512.",
"TakenAlone": "1",
"CounterHTOff": "3"
},
{
"PublicDescription": "Counts all requests that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all requests miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc08fff",
"MSRValue": "0x3FFFC08FFF",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all requests that miss in the L3",
"BriefDescription": "Counts all requests miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x01004007f7",
"MSRValue": "0x01004007F7",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_READS.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram",
"BriefDescription": "miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc007f7",
"MSRValue": "0x3FFFC007F7",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_READS.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3",
"BriefDescription": "miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x0100400244",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch code reads miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00244",
"MSRValue": "0x3FFFC00244",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch code reads that miss in the L3",
"BriefDescription": "Counts all demand & prefetch code reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x0100400122",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch RFOs miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00122",
"MSRValue": "0x3FFFC00122",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch RFOs that miss in the L3",
"BriefDescription": "Counts all demand & prefetch RFOs miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x0100400091",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand & prefetch data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00091",
"MSRValue": "0x3FFFC00091",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch data reads that miss in the L3",
"BriefDescription": "Counts all demand & prefetch data reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00200",
"MSRValue": "0x3FFFC00200",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L3_CODE_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00100",
"MSRValue": "0x3FFFC00100",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L3_RFO.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00080",
"MSRValue": "0x3FFFC00080",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L3_DATA_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads that miss in the L3",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00040",
"MSRValue": "0x3FFFC00040",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads that miss in the L3",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00020",
"MSRValue": "0x3FFFC00020",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs that miss in the L3",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts prefetch (that bring data to L2) data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to L2) data reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00010",
"MSRValue": "0x3FFFC00010",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_DATA_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads that miss in the L3",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand code reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x0100400004",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand code reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand code reads miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00004",
"MSRValue": "0x3FFFC00004",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand code reads that miss in the L3",
"BriefDescription": "Counts all demand code reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x0100400002",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand data writes (RFOs) that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand data writes (RFOs) miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts all demand data writes (RFOs) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00002",
"MSRValue": "0x3FFFC00002",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand data writes (RFOs) that miss in the L3",
"BriefDescription": "Counts all demand data writes (RFOs) miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts demand data reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads miss the L3 and the data is returned from local dram",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x0100400001",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.L3_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts demand data reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts demand data reads miss the L3 and the data is returned from local dram",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "Counts demand data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads miss in the L3",
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x3fffc00001",
"MSRValue": "0x3FFFC00001",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.L3_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"MSRIndex": "0x1a6, 0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts demand data reads that miss in the L3",
"BriefDescription": "Counts demand data reads miss in the L3",
"Offcore": "1",
"CounterHTOff": "0,1,2,3"
}

33
tools/perf/pmu-events/arch/x86/haswell/pipeline.json
View file

@ -1,7 +1,6 @@
[
{
"PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. INST_RETIRED.ANY is counted by a designated fixed counter, leaving the programmable counters available for other events. Faulting executions of GETSEC/VM entry/VM Exit/MWait will not count as retired instructions.",
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"Errata": "HSD140, HSD143",
@ -12,7 +11,6 @@
},
{
"PublicDescription": "This event counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -21,7 +19,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"AnyThread": "1",
@ -32,7 +29,6 @@
},
{
"PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state.",
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",
@ -1071,7 +1067,8 @@
"CounterHTOff": "1"
},
{
"PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
"EventCode": "0xC0",
"Counter": "0,1,2,3",
"UMask": "0x2",
@ -1081,13 +1078,13 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of microcode assists invoked by HW upon uop writeback.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x40",
"EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
"SampleAfterValue": "100003",
"BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -1102,28 +1099,34 @@
"Data_LA": "1"
},
{
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC2",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "UOPS_RETIRED.STALL_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles without actually retired uops.",
"BriefDescription": "Cycles no executable uops retired",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC2",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "UOPS_RETIRED.TOTAL_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with less than 10 actually retired uops.",
"BriefDescription": "Number of cycles using always true condition applied to PEBS uops retired event.",
"CounterMask": "10",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC2",
"Invert": "1",
"Counter": "0,1,2,3",
@ -1131,7 +1134,7 @@
"AnyThread": "1",
"EventName": "UOPS_RETIRED.CORE_STALL_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles without actually retired uops.",
"BriefDescription": "Cycles no executable uops retired on core",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
@ -1245,13 +1248,14 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts the number of not taken branch instructions retired.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC4",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "BR_INST_RETIRED.NOT_TAKEN",
"SampleAfterValue": "400009",
"BriefDescription": "Not taken branch instructions retired.",
"BriefDescription": "Counts all not taken macro branch instructions retired.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -1265,13 +1269,14 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of far branches retired.",
"PEBS": "1",
"PublicDescription": "",
"EventCode": "0xC4",
"Counter": "0,1,2,3",
"UMask": "0x40",
"EventName": "BR_INST_RETIRED.FAR_BRANCH",
"SampleAfterValue": "100003",
"BriefDescription": "Far branch instructions retired.",
"BriefDescription": "Counts the number of far branch instructions retired.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{

173
tools/perf/pmu-events/arch/x86/haswellx/cache.json
View file

@ -64,18 +64,18 @@
},
{
"EventCode": "0x24",
"UMask": "0x41",
"UMask": "0xc1",
"BriefDescription": "Demand Data Read requests that hit L2 cache",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
"Errata": "HSD78",
"PublicDescription": "Demand data read requests that hit L2 cache.",
"PublicDescription": "Counts the number of demand Data Read requests, initiated by load instructions, that hit L2 cache",
"SampleAfterValue": "200003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"UMask": "0x42",
"UMask": "0xc2",
"BriefDescription": "RFO requests that hit L2 cache",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.RFO_HIT",
@ -85,7 +85,7 @@
},
{
"EventCode": "0x24",
"UMask": "0x44",
"UMask": "0xc4",
"BriefDescription": "L2 cache hits when fetching instructions, code reads.",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.CODE_RD_HIT",
@ -95,7 +95,7 @@
},
{
"EventCode": "0x24",
"UMask": "0x50",
"UMask": "0xd0",
"BriefDescription": "L2 prefetch requests that hit L2 cache",
"Counter": "0,1,2,3",
"EventName": "L2_RQSTS.L2_PF_HIT",
@ -416,7 +416,7 @@
{
"EventCode": "0xD0",
"UMask": "0x11",
"BriefDescription": "Retired load uops that miss the STLB. (precise Event)",
"BriefDescription": "Retired load uops that miss the STLB.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -428,7 +428,7 @@
{
"EventCode": "0xD0",
"UMask": "0x12",
"BriefDescription": "Retired store uops that miss the STLB. (precise Event)",
"BriefDescription": "Retired store uops that miss the STLB.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -441,7 +441,7 @@
{
"EventCode": "0xD0",
"UMask": "0x21",
"BriefDescription": "Retired load uops with locked access. (precise Event)",
"BriefDescription": "Retired load uops with locked access.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -453,34 +453,32 @@
{
"EventCode": "0xD0",
"UMask": "0x41",
"BriefDescription": "Retired load uops that split across a cacheline boundary. (precise Event)",
"BriefDescription": "Retired load uops that split across a cacheline boundary.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
"Errata": "HSD29, HSM30",
"PublicDescription": "This event counts load uops retired which had memory addresses spilt across 2 cache lines. A line split is across 64B cache-lines which may include a page split (4K). This is a precise event.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD0",
"UMask": "0x42",
"BriefDescription": "Retired store uops that split across a cacheline boundary. (precise Event)",
"BriefDescription": "Retired store uops that split across a cacheline boundary.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
"Errata": "HSD29, HSM30",
"L1_Hit_Indication": "1",
"PublicDescription": "This event counts store uops retired which had memory addresses spilt across 2 cache lines. A line split is across 64B cache-lines which may include a page split (4K). This is a precise event.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD0",
"UMask": "0x81",
"BriefDescription": "All retired load uops. (precise Event)",
"BriefDescription": "All retired load uops.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -492,14 +490,13 @@
{
"EventCode": "0xD0",
"UMask": "0x82",
"BriefDescription": "All retired store uops. (precise Event)",
"BriefDescription": "All retired store uops.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_UOPS_RETIRED.ALL_STORES",
"Errata": "HSD29, HSM30",
"L1_Hit_Indication": "1",
"PublicDescription": "This event counts all store uops retired. This is a precise event.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3"
},
@ -530,13 +527,13 @@
{
"EventCode": "0xD1",
"UMask": "0x4",
"BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source.",
"BriefDescription": "Retired load uops which data sources were data hits in L3 without snoops required.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L3_HIT",
"Errata": "HSD74, HSD29, HSD25, HSM26, HSM30",
"PublicDescription": "This event counts retired load uops in which data sources were data hits in the L3 cache without snoops required. This does not include hardware prefetches. This is a precise event.",
"PublicDescription": "Retired load uops with L3 cache hits as data sources.",
"SampleAfterValue": "50021",
"CounterHTOff": "0,1,2,3"
},
@ -549,19 +546,20 @@
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
"Errata": "HSM30",
"PublicDescription": "This event counts retired load uops in which data sources missed in the L1 cache. This does not include hardware prefetches. This is a precise event.",
"PublicDescription": "Retired load uops missed L1 cache as data sources.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD1",
"UMask": "0x10",
"BriefDescription": "Retired load uops with L2 cache misses as data sources.",
"BriefDescription": "Miss in mid-level (L2) cache. Excludes Unknown data-source.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
"Errata": "HSD29, HSM30",
"PublicDescription": "Retired load uops missed L2. Unknown data source excluded.",
"SampleAfterValue": "50021",
"CounterHTOff": "0,1,2,3"
},
@ -574,6 +572,7 @@
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_RETIRED.L3_MISS",
"Errata": "HSD74, HSD29, HSD25, HSM26, HSM30",
"PublicDescription": "Retired load uops missed L3. Excludes unknown data source .",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -604,26 +603,24 @@
{
"EventCode": "0xD2",
"UMask": "0x2",
"BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache. ",
"BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT",
"Errata": "HSD29, HSD25, HSM26, HSM30",
"PublicDescription": "This event counts retired load uops that hit in the L3 cache, but required a cross-core snoop which resulted in a HIT in an on-pkg core cache. This does not include hardware prefetches. This is a precise event.",
"SampleAfterValue": "20011",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD2",
"UMask": "0x4",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared L3. ",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared L3.",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM",
"Errata": "HSD29, HSD25, HSM26, HSM30",
"PublicDescription": "This event counts retired load uops that hit in the L3 cache, but required a cross-core snoop which resulted in a HITM (hit modified) in an on-pkg core cache. This does not include hardware prefetches. This is a precise event.",
"SampleAfterValue": "20011",
"CounterHTOff": "0,1,2,3"
},
@ -642,19 +639,20 @@
{
"EventCode": "0xD3",
"UMask": "0x1",
"BriefDescription": "Data from local DRAM either Snoop not needed or Snoop Miss (RspI)",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM",
"Errata": "HSD74, HSD29, HSD25, HSM30",
"PublicDescription": "This event counts retired load uops where the data came from local DRAM. This does not include hardware prefetches. This is a precise event.",
"PublicDescription": "This event counts retired load uops where the data came from local DRAM. This does not include hardware prefetches.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xD3",
"UMask": "0x4",
"BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI) (Precise Event)",
"BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI)",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -666,7 +664,7 @@
{
"EventCode": "0xD3",
"UMask": "0x10",
"BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM (Precise Event)",
"BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -678,7 +676,7 @@
{
"EventCode": "0xD3",
"UMask": "0x20",
"BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache (Precise Event)",
"BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache",
"Data_LA": "1",
"PEBS": "1",
"Counter": "0,1,2,3",
@ -833,7 +831,6 @@
"BriefDescription": "Split locks in SQ",
"Counter": "0,1,2,3",
"EventName": "SQ_MISC.SPLIT_LOCK",
"PublicDescription": "",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -841,12 +838,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts demand data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0001",
"BriefDescription": "Counts demand data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0001",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts demand data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -854,12 +851,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts demand data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0001",
"BriefDescription": "Counts demand data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0001",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts demand data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -867,12 +864,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0002",
"BriefDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -880,12 +877,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0002",
"BriefDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -893,12 +890,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0004",
"BriefDescription": "Counts all demand code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0004",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -906,12 +903,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand code reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0004",
"BriefDescription": "Counts all demand code reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0004",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand code reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -919,12 +916,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3",
"MSRValue": "0x3f803c0010",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads hit in the L3",
"MSRValue": "0x3F803C0010",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_L2_DATA_RD.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to L2) data reads hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -932,12 +929,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3",
"MSRValue": "0x3f803c0020",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs hit in the L3",
"MSRValue": "0x3F803C0020",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -945,12 +942,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3",
"MSRValue": "0x3f803c0040",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads hit in the L3",
"MSRValue": "0x3F803C0040",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -958,12 +955,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3",
"MSRValue": "0x3f803c0080",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads hit in the L3",
"MSRValue": "0x3F803C0080",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_DATA_RD.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -971,12 +968,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3",
"MSRValue": "0x3f803c0100",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs hit in the L3",
"MSRValue": "0x3F803C0100",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_RFO.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -984,12 +981,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3",
"MSRValue": "0x3f803c0200",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads hit in the L3",
"MSRValue": "0x3F803C0200",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_CODE_RD.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -997,12 +994,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0091",
"BriefDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1010,12 +1007,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0091",
"BriefDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1023,12 +1020,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0122",
"BriefDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1036,12 +1033,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c0122",
"BriefDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C0122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1049,12 +1046,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c0244",
"BriefDescription": "Counts all demand & prefetch code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C0244",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1062,12 +1059,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003c07f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"MSRValue": "0x04003C07F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1075,12 +1072,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003c07f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"MSRValue": "0x10003C07F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_HIT.HITM_OTHER_CORE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -1088,12 +1085,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all requests that hit in the L3",
"MSRValue": "0x3f803c8fff",
"BriefDescription": "Counts all requests hit in the L3",
"MSRValue": "0x3F803C8FFF",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.LLC_HIT.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all requests that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all requests hit in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
}

252
tools/perf/pmu-events/arch/x86/haswellx/hsx-metrics.json
View file

@ -1,158 +1,340 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , IDQ.MITE_UOPS / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 16 * ( ICACHE.HIT + ICACHE.MISSES ) / 4.0 ) )",
"MetricGroup": "Frontend",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "( UOPS_EXECUTED.CORE / 2 / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@) ) if #SMT_on else UOPS_EXECUTED.CORE / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "( UOPS_EXECUTED.CORE / 2 / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@) ) if #SMT_on else UOPS_EXECUTED.CORE / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE.IFDATA_STALL - (( 14 * ITLB_MISSES.STLB_HIT + ITLB_MISSES.WALK_DURATION )) ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_UOPS_RETIRED.L1_MISS + mem_load_uops_retired.hit_lfb )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( cpu@l1d_pend_miss.pending_cycles\\,any\\=1@ / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / cycles",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "( 64 * ( uncore_imc@cas_count_read@ + uncore_imc@cas_count_write@ ) / 1000000000 ) / duration_time",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "1000000000 * ( cbox@event\\=0x36\\,umask\\=0x3\\,filter_opc\\=0x182@ / cbox@event\\=0x35\\,umask\\=0x3\\,filter_opc\\=0x182@ ) / ( cbox_0@event\\=0x0@ / duration_time )",
"BriefDescription": "Average latency of data read request to external memory (in nanoseconds). Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_Lat",
"MetricName": "DRAM_Read_Latency"
},
{
"MetricExpr": "cbox@event\\=0x36\\,umask\\=0x3\\,filter_opc\\=0x182@ / cbox@event\\=0x36\\,umask\\=0x3\\,filter_opc\\=0x182\\,thresh\\=1@",
"BriefDescription": "Average number of parallel data read requests to external memory. Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_Parallel_Reads"
},
{
"MetricExpr": "cbox_0@event\\=0x0@",
"BriefDescription": "Socket actual clocks when any core is active on that socket",
"MetricGroup": "",
"MetricName": "Socket_CLKS"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

172
tools/perf/pmu-events/arch/x86/haswellx/memory.json
View file

@ -291,7 +291,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 4.",
"BriefDescription": "Randomly selected loads with latency value being above 4.",
"PEBS": "2",
"MSRValue": "0x4",
"Counter": "3",
@ -305,7 +305,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 8.",
"BriefDescription": "Randomly selected loads with latency value being above 8.",
"PEBS": "2",
"MSRValue": "0x8",
"Counter": "3",
@ -319,7 +319,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 16.",
"BriefDescription": "Randomly selected loads with latency value being above 16.",
"PEBS": "2",
"MSRValue": "0x10",
"Counter": "3",
@ -333,7 +333,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 32.",
"BriefDescription": "Randomly selected loads with latency value being above 32.",
"PEBS": "2",
"MSRValue": "0x20",
"Counter": "3",
@ -347,7 +347,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 64.",
"BriefDescription": "Randomly selected loads with latency value being above 64.",
"PEBS": "2",
"MSRValue": "0x40",
"Counter": "3",
@ -361,7 +361,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 128.",
"BriefDescription": "Randomly selected loads with latency value being above 128.",
"PEBS": "2",
"MSRValue": "0x80",
"Counter": "3",
@ -375,7 +375,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 256.",
"BriefDescription": "Randomly selected loads with latency value being above 256.",
"PEBS": "2",
"MSRValue": "0x100",
"Counter": "3",
@ -389,7 +389,7 @@
{
"EventCode": "0xCD",
"UMask": "0x1",
"BriefDescription": "Loads with latency value being above 512.",
"BriefDescription": "Randomly selected loads with latency value being above 512.",
"PEBS": "2",
"MSRValue": "0x200",
"Counter": "3",
@ -404,12 +404,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts demand data reads that miss in the L3",
"MSRValue": "0x3fbfc00001",
"BriefDescription": "Counts demand data reads miss in the L3",
"MSRValue": "0x3FBFC00001",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts demand data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -417,12 +417,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts demand data reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts demand data reads miss the L3 and the data is returned from local dram",
"MSRValue": "0x0600400001",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts demand data reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts demand data reads miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -430,12 +430,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that miss in the L3",
"MSRValue": "0x3fbfc00002",
"BriefDescription": "Counts all demand data writes (RFOs) miss in the L3",
"MSRValue": "0x3FBFC00002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -443,12 +443,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand data writes (RFOs) miss the L3 and the data is returned from local dram",
"MSRValue": "0x0600400002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -456,12 +456,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand data writes (RFOs) that miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103fc00002",
"BriefDescription": "Counts all demand data writes (RFOs) miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103FC00002",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.REMOTE_HITM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand data writes (RFOs) miss the L3 and the modified data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -469,12 +469,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand code reads that miss in the L3",
"MSRValue": "0x3fbfc00004",
"BriefDescription": "Counts all demand code reads miss in the L3",
"MSRValue": "0x3FBFC00004",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -482,12 +482,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand code reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand code reads miss the L3 and the data is returned from local dram",
"MSRValue": "0x0600400004",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand code reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand code reads miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -495,12 +495,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads that miss in the L3",
"MSRValue": "0x3fbfc00010",
"BriefDescription": "Counts prefetch (that bring data to L2) data reads miss in the L3",
"MSRValue": "0x3FBFC00010",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_L2_DATA_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts prefetch (that bring data to L2) data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to L2) data reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -508,12 +508,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs that miss in the L3",
"MSRValue": "0x3fbfc00020",
"BriefDescription": "Counts all prefetch (that bring data to L2) RFOs miss in the L3",
"MSRValue": "0x3FBFC00020",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to L2) RFOs miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -521,12 +521,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads that miss in the L3",
"MSRValue": "0x3fbfc00040",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) code reads miss in the L3",
"MSRValue": "0x3FBFC00040",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -534,12 +534,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads that miss in the L3",
"MSRValue": "0x3fbfc00080",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) data reads miss in the L3",
"MSRValue": "0x3FBFC00080",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_DATA_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -547,12 +547,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3",
"MSRValue": "0x3fbfc00100",
"BriefDescription": "Counts all prefetch (that bring data to LLC only) RFOs miss in the L3",
"MSRValue": "0x3FBFC00100",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -560,12 +560,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3",
"MSRValue": "0x3fbfc00200",
"BriefDescription": "Counts prefetch (that bring data to LLC only) code reads miss in the L3",
"MSRValue": "0x3FBFC00200",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.PF_LLC_CODE_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts prefetch (that bring data to LLC only) code reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -573,12 +573,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss in the L3",
"MSRValue": "0x3fbfc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss in the L3",
"MSRValue": "0x3FBFC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -586,12 +586,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from local dram",
"MSRValue": "0x0600400091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -599,12 +599,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063f800091",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063F800091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from remote dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the data is returned from remote dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -612,12 +612,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103fc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103FC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_HITM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and the modified data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -625,12 +625,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch data reads that miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x083fc00091",
"BriefDescription": "Counts all demand & prefetch data reads miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x083FC00091",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_HIT_FORWARD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and clean or shared data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch data reads miss the L3 and clean or shared data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -638,12 +638,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that miss in the L3",
"MSRValue": "0x3fbfc00122",
"BriefDescription": "Counts all demand & prefetch RFOs miss in the L3",
"MSRValue": "0x3FBFC00122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -651,12 +651,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch RFOs miss the L3 and the data is returned from local dram",
"MSRValue": "0x0600400122",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch RFOs miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -664,12 +664,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch code reads that miss in the L3",
"MSRValue": "0x3fbfc00244",
"BriefDescription": "Counts all demand & prefetch code reads miss in the L3",
"MSRValue": "0x3FBFC00244",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -677,12 +677,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram",
"BriefDescription": "Counts all demand & prefetch code reads miss the L3 and the data is returned from local dram",
"MSRValue": "0x0600400244",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all demand & prefetch code reads miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -690,12 +690,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3",
"MSRValue": "0x3fbfc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss in the L3",
"MSRValue": "0x3FBFC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -703,12 +703,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram",
"MSRValue": "0x06004007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from local dram",
"MSRValue": "0x06004007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.LOCAL_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from local dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -716,12 +716,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063f8007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from remote dram",
"MSRValue": "0x063F8007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_DRAM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from remote dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the data is returned from remote dram",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -729,12 +729,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103fc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the modified data is transferred from remote cache",
"MSRValue": "0x103FC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_HITM",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and the modified data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -742,12 +742,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x083fc007f7",
"BriefDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and clean or shared data is transferred from remote cache",
"MSRValue": "0x083FC007F7",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_HIT_FORWARD",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and clean or shared data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) miss the L3 and clean or shared data is transferred from remote cache",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
},
@ -755,12 +755,12 @@
"Offcore": "1",
"EventCode": "0xB7, 0xBB",
"UMask": "0x1",
"BriefDescription": "Counts all requests that miss in the L3",
"MSRValue": "0x3fbfc08fff",
"BriefDescription": "Counts all requests miss in the L3",
"MSRValue": "0x3FBFC08FFF",
"Counter": "0,1,2,3",
"EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.LLC_MISS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"PublicDescription": "Counts all requests that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
"PublicDescription": "Counts all requests miss in the L3",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3"
}

10
tools/perf/pmu-events/arch/x86/haswellx/pipeline.json
View file

@ -1,6 +1,5 @@
[
{
"EventCode": "0x00",
"UMask": "0x1",
"BriefDescription": "Instructions retired from execution.",
"Counter": "Fixed counter 0",
@ -11,7 +10,6 @@
"CounterHTOff": "Fixed counter 0"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when the thread is not in halt state.",
"Counter": "Fixed counter 1",
@ -21,7 +19,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
"Counter": "Fixed counter 1",
@ -31,7 +28,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x3",
"BriefDescription": "Reference cycles when the core is not in halt state.",
"Counter": "Fixed counter 2",
@ -1098,6 +1094,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.ALL",
"PublicDescription": "Counts the number of micro-ops retired. Use Cmask=1 and invert to count active cycles or stalled cycles.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1142,6 +1139,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.RETIRE_SLOTS",
"PublicDescription": "This event counts the number of retirement slots used each cycle. There are potentially 4 slots that can be used each cycle - meaning, 4 uops or 4 instructions could retire each cycle.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1201,6 +1199,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.CONDITIONAL",
"PublicDescription": "Counts the number of conditional branch instructions retired.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1241,6 +1240,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NEAR_RETURN",
"PublicDescription": "Counts the number of near return instructions retired.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1261,6 +1261,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NEAR_TAKEN",
"PublicDescription": "Number of near taken branches retired.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -1312,6 +1313,7 @@
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
"PublicDescription": "Number of near branch instructions retired that were taken but mispredicted.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},

10
tools/perf/pmu-events/arch/x86/ivybridge/cache.json
View file

@ -1012,7 +1012,7 @@
"EventName": "OFFCORE_RESPONSE.SPLIT_LOCK_UC_LOCK.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts requests where the address of an atomic lock instruction spans a cache line boundary or the lock instruction is executed on uncacheable address ",
"BriefDescription": "Counts requests where the address of an atomic lock instruction spans a cache line boundary or the lock instruction is executed on uncacheable address",
"CounterHTOff": "0,1,2,3"
},
{
@ -1036,7 +1036,7 @@
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand data reads ",
"BriefDescription": "Counts all demand data reads",
"CounterHTOff": "0,1,2,3"
},
{
@ -1048,7 +1048,7 @@
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand rfo's ",
"BriefDescription": "Counts all demand rfo's",
"CounterHTOff": "0,1,2,3"
},
{
@ -1084,7 +1084,7 @@
"EventName": "OFFCORE_RESPONSE.ALL_RFO.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all demand & prefetch prefetch RFOs ",
"BriefDescription": "Counts all demand & prefetch prefetch RFOs",
"CounterHTOff": "0,1,2,3"
},
{
@ -1096,7 +1096,7 @@
"EventName": "OFFCORE_RESPONSE.ALL_READS.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": "Counts all data/code/rfo references (demand & prefetch) ",
"BriefDescription": "Counts all data/code/rfo references (demand & prefetch)",
"CounterHTOff": "0,1,2,3"
}
]

250
tools/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json
View file

@ -1,164 +1,340 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4) )",
"MetricGroup": "Frontend",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE.IFETCH_STALL ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_UOPS_RETIRED.L1_MISS + mem_load_uops_retired.hit_lfb )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( cpu@l1d_pend_miss.pending_cycles\\,any\\=1@ / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / cycles",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.LLC_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2* FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4*( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8* SIMD_FP_256.PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "64 * ( arb@event\\=0x81\\,umask\\=0x1@ + arb@event\\=0x84\\,umask\\=0x1@ ) / 1000000 / duration_time / 1000",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

4
tools/perf/pmu-events/arch/x86/ivybridge/pipeline.json
View file

@ -1,6 +1,5 @@
[
{
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -9,7 +8,6 @@
"CounterHTOff": "Fixed counter 0"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -19,7 +17,6 @@
},
{
"PublicDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"AnyThread": "1",
@ -29,7 +26,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",

256
tools/perf/pmu-events/arch/x86/ivytown/ivt-metrics.json
View file

@ -1,164 +1,346 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4) )",
"MetricGroup": "Frontend",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_UOPS_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( cpu@UOPS_EXECUTED.CORE\\,cmask\\=1@ / 2) if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE.IFETCH_STALL ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_UOPS_RETIRED.L1_MISS + mem_load_uops_retired.hit_lfb )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( cpu@l1d_pend_miss.pending_cycles\\,any\\=1@ / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / cycles",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_DURATION + DTLB_LOAD_MISSES.WALK_DURATION + DTLB_STORE_MISSES.WALK_DURATION ) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_UOPS_RETIRED.LLC_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2* FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4*( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8* SIMD_FP_256.PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "( 64 * ( uncore_imc@cas_count_read@ + uncore_imc@cas_count_write@ ) / 1000000000 ) / duration_time",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "cbox_0@event\\=0x0@",
"BriefDescription": "Socket actual clocks when any core is active on that socket",
"MetricGroup": "",
"MetricName": "Socket_CLKS"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

4
tools/perf/pmu-events/arch/x86/ivytown/pipeline.json
View file

@ -1,6 +1,5 @@
[
{
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -9,7 +8,6 @@
"CounterHTOff": "Fixed counter 0"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -19,7 +17,6 @@
},
{
"PublicDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"AnyThread": "1",
@ -29,7 +26,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",

6
tools/perf/pmu-events/arch/x86/jaketown/cache.json
View file

@ -31,7 +31,7 @@
},
{
"PEBS": "1",
"PublicDescription": "This event counts line-split load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"PublicDescription": "This event counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x41",
@ -42,7 +42,7 @@
},
{
"PEBS": "1",
"PublicDescription": "This event counts line-split store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"PublicDescription": "This event counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x42",
@ -179,7 +179,7 @@
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "This event counts L1D data line replacements. Replacements occur when a new line is brought into the cache, causing eviction of a line loaded earlier. ",
"PublicDescription": "This event counts L1D data line replacements. Replacements occur when a new line is brought into the cache, causing eviction of a line loaded earlier.",
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x1",

150
tools/perf/pmu-events/arch/x86/jaketown/jkt-metrics.json
View file

@ -1,140 +1,232 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4) )",
"MetricGroup": "Frontend",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_DISPATCHED.THREAD / (( cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_DISPATCHED.THREAD / (( cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@ / 2) if #SMT_on else cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2* FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4*( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8* SIMD_FP_256.PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "( 64 * ( uncore_imc@cas_count_read@ + uncore_imc@cas_count_write@ ) / 1000000000 ) / duration_time",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "cbox_0@event\\=0x0@",
"BriefDescription": "Socket actual clocks when any core is active on that socket",
"MetricGroup": "",
"MetricName": "Socket_CLKS"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

12
tools/perf/pmu-events/arch/x86/jaketown/pipeline.json
View file

@ -1,7 +1,6 @@
[
{
"PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. ",
"EventCode": "0x00",
"PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers.",
"Counter": "Fixed counter 1",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -10,8 +9,7 @@
"CounterHTOff": "Fixed counter 1"
},
{
"PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. ",
"EventCode": "0x00",
"PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
"Counter": "Fixed counter 2",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -20,8 +18,7 @@
"CounterHTOff": "Fixed counter 2"
},
{
"PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. ",
"EventCode": "0x00",
"PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
"Counter": "Fixed counter 3",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",
@ -778,7 +775,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts loads that followed a store to the same address, where the data could not be forwarded inside the pipeline from the store to the load. The most common reason why store forwarding would be blocked is when a load's address range overlaps with a preceding smaller uncompleted store. See the table of not supported store forwards in the Intel? 64 and IA-32 Architectures Optimization Reference Manual. The penalty for blocked store forwarding is that the load must wait for the store to complete before it can be issued.",
"PublicDescription": "This event counts loads that followed a store to the same address, where the data could not be forwarded inside the pipeline from the store to the load. The most common reason why store forwarding would be blocked is when a load's address range overlaps with a preceeding smaller uncompleted store. See the table of not supported store forwards in the Intel? 64 and IA-32 Architectures Optimization Reference Manual. The penalty for blocked store forwarding is that the load must wait for the store to complete before it can be issued.",
"EventCode": "0x03",
"Counter": "0,1,2,3",
"UMask": "0x2",
@ -1098,7 +1095,6 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x2",
"AnyThread": "1",

666
tools/perf/pmu-events/arch/x86/knightslanding/cache.json
View file

File diff suppressed because it is too large Load diff

268
tools/perf/pmu-events/arch/x86/knightslanding/memory.json
View file

@ -9,18 +9,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400070 ",
"MSRValue": "0x0100400070",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_PF_L2.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any Prefetch requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts any Prefetch requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200070 ",
"MSRValue": "0x0080200070",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_PF_L2.MCDRAM_NEAR",
@ -31,18 +31,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000070 ",
"MSRValue": "0x0101000070",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_PF_L2.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any Prefetch requests that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts any Prefetch requests that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800070 ",
"MSRValue": "0x0080800070",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_PF_L2.DDR_NEAR",
@ -53,18 +53,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x01004032f7 ",
"MSRValue": "0x01004032f7",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any Read request that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts any Read request that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x00802032f7 ",
"MSRValue": "0x00802032f7",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.MCDRAM_NEAR",
@ -75,18 +75,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x01010032f7 ",
"MSRValue": "0x01010032f7",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any Read request that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts any Read request that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x00808032f7 ",
"MSRValue": "0x00808032f7",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.DDR_NEAR",
@ -97,18 +97,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400044 ",
"MSRValue": "0x0100400044",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_CODE_RD.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand code reads and prefetch code read requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Demand code reads and prefetch code read requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200044 ",
"MSRValue": "0x0080200044",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_CODE_RD.MCDRAM_NEAR",
@ -119,18 +119,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000044 ",
"MSRValue": "0x0101000044",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_CODE_RD.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand code reads and prefetch code read requests that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Demand code reads and prefetch code read requests that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800044 ",
"MSRValue": "0x0080800044",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_CODE_RD.DDR_NEAR",
@ -141,18 +141,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400022 ",
"MSRValue": "0x0100400022",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand cacheable data write requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Demand cacheable data write requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200022 ",
"MSRValue": "0x0080200022",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.MCDRAM_NEAR",
@ -163,18 +163,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000022 ",
"MSRValue": "0x0101000022",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand cacheable data write requests that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Demand cacheable data write requests that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800022 ",
"MSRValue": "0x0080800022",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.DDR_NEAR",
@ -185,18 +185,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100403091 ",
"MSRValue": "0x0100403091",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand cacheable data and L1 prefetch data read requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Demand cacheable data and L1 prefetch data read requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080203091 ",
"MSRValue": "0x0080203091",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.MCDRAM_NEAR",
@ -207,18 +207,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101003091 ",
"MSRValue": "0x0101003091",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand cacheable data and L1 prefetch data read requests that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Demand cacheable data and L1 prefetch data read requests that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080803091 ",
"MSRValue": "0x0080803091",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.DDR_NEAR",
@ -229,18 +229,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100408000 ",
"MSRValue": "0x0100408000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any request that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts any request that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080208000 ",
"MSRValue": "0x0080208000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.MCDRAM_NEAR",
@ -251,18 +251,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101008000 ",
"MSRValue": "0x0101008000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts any request that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts any request that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080808000 ",
"MSRValue": "0x0080808000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.DDR_NEAR",
@ -273,18 +273,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100402000 ",
"MSRValue": "0x0100402000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L1_DATA_RD.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts L1 data HW prefetches that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts L1 data HW prefetches that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080202000 ",
"MSRValue": "0x0080202000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L1_DATA_RD.MCDRAM_NEAR",
@ -295,18 +295,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101002000 ",
"MSRValue": "0x0101002000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L1_DATA_RD.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts L1 data HW prefetches that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts L1 data HW prefetches that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080802000 ",
"MSRValue": "0x0080802000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L1_DATA_RD.DDR_NEAR",
@ -317,18 +317,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100401000 ",
"MSRValue": "0x0100401000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_SOFTWARE.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Software Prefetches that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Software Prefetches that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080201000 ",
"MSRValue": "0x0080201000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_SOFTWARE.MCDRAM_NEAR",
@ -339,18 +339,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101001000 ",
"MSRValue": "0x0101001000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_SOFTWARE.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Software Prefetches that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Software Prefetches that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080801000 ",
"MSRValue": "0x0080801000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_SOFTWARE.DDR_NEAR",
@ -361,18 +361,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400400 ",
"MSRValue": "0x0100400400",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Bus locks and split lock requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Bus locks and split lock requests that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200400 ",
"MSRValue": "0x0080200400",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.MCDRAM_NEAR",
@ -383,18 +383,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000400 ",
"MSRValue": "0x0101000400",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Bus locks and split lock requests that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Bus locks and split lock requests that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800400 ",
"MSRValue": "0x0080800400",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.DDR_NEAR",
@ -405,18 +405,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400200 ",
"MSRValue": "0x0100400200",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_READS.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts UC code reads (valid only for Outstanding response type) that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts UC code reads (valid only for Outstanding response type) that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200200 ",
"MSRValue": "0x0080200200",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_READS.MCDRAM_NEAR",
@ -427,18 +427,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000200 ",
"MSRValue": "0x0101000200",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_READS.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts UC code reads (valid only for Outstanding response type) that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts UC code reads (valid only for Outstanding response type) that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800200 ",
"MSRValue": "0x0080800200",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_READS.DDR_NEAR",
@ -449,18 +449,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400100 ",
"MSRValue": "0x0100400100",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_WRITES.MCDRAM_FAR",
"MSRIndex": "0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Partial writes (UC or WT or WP and should be programmed on PMC1) that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Partial writes (UC or WT or WP and should be programmed on PMC1) that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200100 ",
"MSRValue": "0x0080200100",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_WRITES.MCDRAM_NEAR",
@ -471,18 +471,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000100 ",
"MSRValue": "0x0101000100",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_WRITES.DDR_FAR",
"MSRIndex": "0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Partial writes (UC or WT or WP and should be programmed on PMC1) that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Partial writes (UC or WT or WP and should be programmed on PMC1) that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800100 ",
"MSRValue": "0x0080800100",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_WRITES.DDR_NEAR",
@ -493,7 +493,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x2000020080 ",
"MSRValue": "0x2000020080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.NON_DRAM",
@ -504,18 +504,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400080 ",
"MSRValue": "0x0100400080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Partial reads (UC or WC and is valid only for Outstanding response type). that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Partial reads (UC or WC and is valid only for Outstanding response type). that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200080 ",
"MSRValue": "0x0080200080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.MCDRAM_NEAR",
@ -526,18 +526,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000080 ",
"MSRValue": "0x0101000080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Partial reads (UC or WC and is valid only for Outstanding response type). that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Partial reads (UC or WC and is valid only for Outstanding response type). that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800080 ",
"MSRValue": "0x0080800080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.DDR_NEAR",
@ -548,18 +548,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400040 ",
"MSRValue": "0x0100400040",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts L2 code HW prefetches that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts L2 code HW prefetches that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200040 ",
"MSRValue": "0x0080200040",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.MCDRAM_NEAR",
@ -570,18 +570,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000040 ",
"MSRValue": "0x0101000040",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts L2 code HW prefetches that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts L2 code HW prefetches that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800040 ",
"MSRValue": "0x0080800040",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.DDR_NEAR",
@ -592,7 +592,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x2000020020 ",
"MSRValue": "0x2000020020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.NON_DRAM",
@ -603,18 +603,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400020 ",
"MSRValue": "0x0100400020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts L2 data RFO prefetches (includes PREFETCHW instruction) that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts L2 data RFO prefetches (includes PREFETCHW instruction) that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200020 ",
"MSRValue": "0x0080200020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.MCDRAM_NEAR",
@ -625,18 +625,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000020 ",
"MSRValue": "0x0101000020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts L2 data RFO prefetches (includes PREFETCHW instruction) that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts L2 data RFO prefetches (includes PREFETCHW instruction) that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800020 ",
"MSRValue": "0x0080800020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.DDR_NEAR",
@ -647,18 +647,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400004 ",
"MSRValue": "0x0100400004",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand code reads and prefetch code reads that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts demand code reads and prefetch code reads that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200004 ",
"MSRValue": "0x0080200004",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.MCDRAM_NEAR",
@ -669,18 +669,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000004 ",
"MSRValue": "0x0101000004",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand code reads and prefetch code reads that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts demand code reads and prefetch code reads that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800004 ",
"MSRValue": "0x0080800004",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.DDR_NEAR",
@ -691,18 +691,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400002 ",
"MSRValue": "0x0100400002",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand cacheable data writes that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts Demand cacheable data writes that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200002 ",
"MSRValue": "0x0080200002",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.MCDRAM_NEAR",
@ -713,18 +713,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000002 ",
"MSRValue": "0x0101000002",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts Demand cacheable data writes that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts Demand cacheable data writes that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800002 ",
"MSRValue": "0x0080800002",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.DDR_NEAR",
@ -735,18 +735,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0100400001 ",
"MSRValue": "0x0100400001",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.MCDRAM_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand cacheable data and L1 prefetch data reads that accounts for data responses from MCDRAM Far or Other tile L2 hit far. ",
"BriefDescription": "Counts demand cacheable data and L1 prefetch data reads that accounts for data responses from MCDRAM Far or Other tile L2 hit far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080200001 ",
"MSRValue": "0x0080200001",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.MCDRAM_NEAR",
@ -757,18 +757,18 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0101000001 ",
"MSRValue": "0x0101000001",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.DDR_FAR",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100007",
"BriefDescription": "Counts demand cacheable data and L1 prefetch data reads that accounts for data responses from DRAM Far. ",
"BriefDescription": "Counts demand cacheable data and L1 prefetch data reads that accounts for data responses from DRAM Far.",
"Offcore": "1"
},
{
"EventCode": "0xB7",
"MSRValue": "0x0080800001 ",
"MSRValue": "0x0080800001",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.DDR_NEAR",
@ -779,7 +779,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600001 ",
"MSRValue": "0x0180600001",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.MCDRAM",
@ -790,7 +790,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600002 ",
"MSRValue": "0x0180600002",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.MCDRAM",
@ -801,7 +801,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600004 ",
"MSRValue": "0x0180600004",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.MCDRAM",
@ -812,7 +812,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600020 ",
"MSRValue": "0x0180600020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.MCDRAM",
@ -823,7 +823,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600080 ",
"MSRValue": "0x0180600080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.MCDRAM",
@ -834,7 +834,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600100 ",
"MSRValue": "0x0180600100",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_WRITES.MCDRAM",
@ -845,7 +845,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600200 ",
"MSRValue": "0x0180600200",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_READS.MCDRAM",
@ -856,7 +856,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600400 ",
"MSRValue": "0x0180600400",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.MCDRAM",
@ -867,7 +867,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180601000 ",
"MSRValue": "0x0180601000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_SOFTWARE.MCDRAM",
@ -878,7 +878,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180608000 ",
"MSRValue": "0x0180608000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.MCDRAM",
@ -889,7 +889,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180603091 ",
"MSRValue": "0x0180603091",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.MCDRAM",
@ -900,7 +900,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600022 ",
"MSRValue": "0x0180600022",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.MCDRAM",
@ -911,7 +911,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600044 ",
"MSRValue": "0x0180600044",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_CODE_RD.MCDRAM",
@ -922,7 +922,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x01806032f7 ",
"MSRValue": "0x01806032f7",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.MCDRAM",
@ -933,7 +933,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0180600070 ",
"MSRValue": "0x0180600070",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_PF_L2.MCDRAM",
@ -944,7 +944,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800001 ",
"MSRValue": "0x0181800001",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_DATA_RD.DDR",
@ -955,7 +955,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800002 ",
"MSRValue": "0x0181800002",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.DDR",
@ -966,7 +966,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800004 ",
"MSRValue": "0x0181800004",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.DEMAND_CODE_RD.DDR",
@ -977,7 +977,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800020 ",
"MSRValue": "0x0181800020",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_RFO.DDR",
@ -988,7 +988,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800040 ",
"MSRValue": "0x0181800040",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L2_CODE_RD.DDR",
@ -999,7 +999,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800080 ",
"MSRValue": "0x0181800080",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PARTIAL_READS.DDR",
@ -1010,7 +1010,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800200 ",
"MSRValue": "0x0181800200",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.UC_CODE_READS.DDR",
@ -1021,7 +1021,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800400 ",
"MSRValue": "0x0181800400",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.BUS_LOCKS.DDR",
@ -1032,7 +1032,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181801000 ",
"MSRValue": "0x0181801000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_SOFTWARE.DDR",
@ -1043,7 +1043,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181802000 ",
"MSRValue": "0x0181802000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.PF_L1_DATA_RD.DDR",
@ -1054,7 +1054,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181808000 ",
"MSRValue": "0x0181808000",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.DDR",
@ -1065,7 +1065,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181803091 ",
"MSRValue": "0x0181803091",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.DDR",
@ -1076,7 +1076,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800022 ",
"MSRValue": "0x0181800022",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_RFO.DDR",
@ -1087,7 +1087,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x0181800044 ",
"MSRValue": "0x0181800044",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_CODE_RD.DDR",
@ -1098,7 +1098,7 @@
},
{
"EventCode": "0xB7",
"MSRValue": "0x01818032f7 ",
"MSRValue": "0x01818032f7",
"Counter": "0,1",
"UMask": "0x1",
"EventName": "OFFCORE_RESPONSE.ANY_READ.DDR",

15
tools/perf/pmu-events/arch/x86/knightslanding/pipeline.json
View file

@ -144,7 +144,7 @@
"BriefDescription": "Counts the number of micro-ops retired that are from the complex flows issued by the micro-sequencer (MS)."
},
{
"PublicDescription": "This event counts the number of micro-ops (uops) retired. The processor decodes complex macro instructions into a sequence of simpler uops. Most instructions are composed of one or two uops. Some instructions are decoded into longer sequences such as repeat instructions, floating point transcendental instructions, and assists. ",
"PublicDescription": "This event counts the number of micro-ops (uops) retired. The processor decodes complex macro instructions into a sequence of simpler uops. Most instructions are composed of one or two uops. Some instructions are decoded into longer sequences such as repeat instructions, floating point transcendental instructions, and assists.",
"EventCode": "0xC2",
"Counter": "0,1",
"UMask": "0x10",
@ -218,7 +218,7 @@
"UMask": "0x20",
"EventName": "NO_ALLOC_CYCLES.RAT_STALL",
"SampleAfterValue": "200003",
"BriefDescription": "Counts the number of core cycles when no micro-ops are allocated and a RATstall (caused by reservation station full) is asserted. "
"BriefDescription": "Counts the number of core cycles when no micro-ops are allocated and a RATstall (caused by reservation station full) is asserted."
},
{
"PublicDescription": "This event counts the number of core cycles when no uops are allocated, the instruction queue is empty and the alloc pipe is stalled waiting for instructions to be fetched.",
@ -251,7 +251,7 @@
"UMask": "0x1f",
"EventName": "RS_FULL_STALL.ALL",
"SampleAfterValue": "200003",
"BriefDescription": "Counts the total number of core cycles the Alloc pipeline is stalled when any one of the reservation stations is full. "
"BriefDescription": "Counts the total number of core cycles the Alloc pipeline is stalled when any one of the reservation stations is full."
},
{
"EventCode": "0xC0",
@ -268,11 +268,10 @@
"UMask": "0x1",
"EventName": "CYCLES_DIV_BUSY.ALL",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles the number of core cycles when divider is busy. Does not imply a stall waiting for the divider. "
"BriefDescription": "Cycles the number of core cycles when divider is busy. Does not imply a stall waiting for the divider."
},
{
"PublicDescription": "This event counts the number of instructions that retire. For instructions that consist of multiple micro-ops, this event counts exactly once, as the last micro-op of the instruction retires. The event continues counting while instructions retire, including during interrupt service routines caused by hardware interrupts, faults or traps.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -296,8 +295,7 @@
"BriefDescription": "Counts the number of unhalted reference clock cycles"
},
{
"PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter\r\n",
"EventCode": "0x00",
"PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter",
"Counter": "Fixed counter 2",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -305,7 +303,6 @@
"BriefDescription": "Fixed Counter: Counts the number of unhalted core clock cycles"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 3",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",
@ -343,7 +340,7 @@
"UMask": "0x1",
"EventName": "RECYCLEQ.LD_BLOCK_ST_FORWARD",
"SampleAfterValue": "200003",
"BriefDescription": "Counts the number of occurences a retired load gets blocked because its address partially overlaps with a store ",
"BriefDescription": "Counts the number of occurences a retired load gets blocked because its address partially overlaps with a store",
"Data_LA": "1"
},
{

2
tools/perf/pmu-events/arch/x86/knightslanding/virtual-memory.json
View file

@ -36,7 +36,7 @@
"EdgeDetect": "1"
},
{
"PublicDescription": "This event counts every cycle when an I-side (walks due to an instruction fetch) page walk is in progress. ",
"PublicDescription": "This event counts every cycle when an I-side (walks due to an instruction fetch) page walk is in progress.",
"EventCode": "0x05",
"Counter": "0,1",
"UMask": "0x2",

870
tools/perf/pmu-events/arch/x86/sandybridge/cache.json
View file

@ -1,329 +1,4 @@
[
{
"PEBS": "1",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x11",
"EventName": "MEM_UOPS_RETIRED.STLB_MISS_LOADS",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops that miss the STLB.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x12",
"EventName": "MEM_UOPS_RETIRED.STLB_MISS_STORES",
"SampleAfterValue": "100003",
"BriefDescription": "Retired store uops that miss the STLB.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x21",
"EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
"SampleAfterValue": "100007",
"BriefDescription": "Retired load uops with locked access.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts line-split load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x41",
"EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops that split across a cacheline boundary.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts line-split store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x42",
"EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
"SampleAfterValue": "100003",
"BriefDescription": "Retired store uops that split across a cacheline boundary.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts the number of load uops retired",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x81",
"EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
"SampleAfterValue": "2000003",
"BriefDescription": "All retired load uops.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts the number of store uops retired.",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x82",
"EventName": "MEM_UOPS_RETIRED.ALL_STORES",
"SampleAfterValue": "2000003",
"BriefDescription": "All retired store uops.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
"SampleAfterValue": "2000003",
"BriefDescription": "Retired load uops with L1 cache hits as data sources.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops with L2 cache hits as data sources.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired load uops that hit in the last-level (L3) cache without snoops required.",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "MEM_LOAD_UOPS_RETIRED.LLC_HIT",
"SampleAfterValue": "50021",
"BriefDescription": "Retired load uops which data sources were data hits in LLC without snoops required.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x40",
"EventName": "MEM_LOAD_UOPS_RETIRED.HIT_LFB",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_MISS",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were LLC hit and cross-core snoop missed in on-pkg core cache.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired load uops that hit in the last-level cache (L3) and were found in a non-modified state in a neighboring core's private cache (same package). Since the last level cache is inclusive, hits to the L3 may require snooping the private L2 caches of any cores on the same socket that have the line. In this case, a snoop was required, and another L2 had the line in a non-modified state.",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HIT",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were LLC and cross-core snoop hits in on-pkg core cache.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired load uops that hit in the last-level cache (L3) and were found in a non-modified state in a neighboring core's private cache (same package). Since the last level cache is inclusive, hits to the L3 may require snooping the private L2 caches of any cores on the same socket that have the line. In this case, a snoop was required, and another L2 had the line in a modified state, so the line had to be invalidated in that L2 cache and transferred to the requesting L2.",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HITM",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared LLC.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_NONE",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops which data sources were hits in LLC without snoops required.",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired demand loads that missed the last-level (L3) cache. This means that the load is usually satisfied from memory in a client system or possibly from the remote socket in a server. Demand loads are non speculative load uops.",
"EventCode": "0xD4",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MEM_LOAD_UOPS_MISC_RETIRED.LLC_MISS",
"SampleAfterValue": "100007",
"BriefDescription": "Retired load uops with unknown information as data source in cache serviced the load.",
"CounterHTOff": "0,1,2,3"
},
{
"PublicDescription": "This event counts L1D data line replacements. Replacements occur when a new line is brought into the cache, causing eviction of a line loaded earlier. ",
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "L1D.REPLACEMENT",
"SampleAfterValue": "2000003",
"BriefDescription": "L1D data line replacements.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "L1D.ALLOCATED_IN_M",
"SampleAfterValue": "2000003",
"BriefDescription": "Allocated L1D data cache lines in M state.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "L1D.EVICTION",
"SampleAfterValue": "2000003",
"BriefDescription": "L1D data cache lines in M state evicted due to replacement.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "L1D.ALL_M_REPLACEMENT",
"SampleAfterValue": "2000003",
"BriefDescription": "Cache lines in M state evicted out of L1D due to Snoop HitM or dirty line replacement.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x48",
"Counter": "2",
"UMask": "0x1",
"EventName": "L1D_PEND_MISS.PENDING",
"SampleAfterValue": "2000003",
"BriefDescription": "L1D miss oustandings duration in cycles.",
"CounterHTOff": "2"
},
{
"EventCode": "0x48",
"Counter": "2",
"UMask": "0x1",
"EventName": "L1D_PEND_MISS.PENDING_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with L1D load Misses outstanding.",
"CounterMask": "1",
"CounterHTOff": "2"
},
{
"EventCode": "0x63",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "LOCK_CYCLES.CACHE_LOCK_DURATION",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when L1D is locked.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding Demand Data Read transactions in uncore queue.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_RFO",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding RFO store transactions in SuperQueue (SQ), queue to uncore.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when offcore outstanding cacheable Core Data Read transactions are present in SuperQueue (SQ), queue to uncore.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD",
"SampleAfterValue": "100003",
"BriefDescription": "Demand Data Read requests sent to uncore.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD",
"SampleAfterValue": "100003",
"BriefDescription": "Cacheable and noncachaeble code read requests.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "OFFCORE_REQUESTS.DEMAND_RFO",
"SampleAfterValue": "100003",
"BriefDescription": "Demand RFO requests including regular RFOs, locks, ItoM.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OFFCORE_REQUESTS.ALL_DATA_RD",
"SampleAfterValue": "100003",
"BriefDescription": "Demand and prefetch data reads.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB2",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_BUFFER.SQ_FULL",
"SampleAfterValue": "2000003",
"BriefDescription": "Cases when offcore requests buffer cannot take more entries for core.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
@ -333,6 +8,15 @@
"BriefDescription": "Demand Data Read requests that hit L2 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x3",
"EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD",
"SampleAfterValue": "200003",
"BriefDescription": "Demand Data Read requests.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
@ -351,6 +35,15 @@
"BriefDescription": "RFO requests that miss L2 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0xc",
"EventName": "L2_RQSTS.ALL_RFO",
"SampleAfterValue": "200003",
"BriefDescription": "RFO requests to L2 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
@ -369,6 +62,15 @@
"BriefDescription": "L2 cache misses when fetching instructions.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x30",
"EventName": "L2_RQSTS.ALL_CODE_RD",
"SampleAfterValue": "200003",
"BriefDescription": "L2 code requests.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
@ -387,6 +89,15 @@
"BriefDescription": "Requests from the L2 hardware prefetchers that miss L2 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0xc0",
"EventName": "L2_RQSTS.ALL_PF",
"SampleAfterValue": "200003",
"BriefDescription": "Requests from L2 hardware prefetchers.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x27",
"Counter": "0,1,2,3",
@ -468,6 +179,400 @@
"BriefDescription": "Not rejected writebacks from L1D to L2 cache lines in any state.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x2E",
"Counter": "0,1,2,3",
"UMask": "0x41",
"EventName": "LONGEST_LAT_CACHE.MISS",
"SampleAfterValue": "100003",
"BriefDescription": "Core-originated cacheable demand requests missed LLC.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x2E",
"Counter": "0,1,2,3",
"UMask": "0x4f",
"EventName": "LONGEST_LAT_CACHE.REFERENCE",
"SampleAfterValue": "100003",
"BriefDescription": "Core-originated cacheable demand requests that refer to LLC.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x48",
"Counter": "2",
"UMask": "0x1",
"EventName": "L1D_PEND_MISS.PENDING",
"SampleAfterValue": "2000003",
"BriefDescription": "L1D miss oustandings duration in cycles.",
"CounterHTOff": "2"
},
{
"EventCode": "0x48",
"Counter": "2",
"UMask": "0x1",
"EventName": "L1D_PEND_MISS.PENDING_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with L1D load Misses outstanding.",
"CounterMask": "1",
"CounterHTOff": "2"
},
{
"EventCode": "0x48",
"Counter": "2",
"UMask": "0x1",
"AnyThread": "1",
"EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
"CounterMask": "1",
"CounterHTOff": "2"
},
{
"EventCode": "0x48",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "L1D_PEND_MISS.FB_FULL",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts L1D data line replacements. Replacements occur when a new line is brought into the cache, causing eviction of a line loaded earlier.",
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "L1D.REPLACEMENT",
"SampleAfterValue": "2000003",
"BriefDescription": "L1D data line replacements.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "L1D.ALLOCATED_IN_M",
"SampleAfterValue": "2000003",
"BriefDescription": "Allocated L1D data cache lines in M state.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "L1D.EVICTION",
"SampleAfterValue": "2000003",
"BriefDescription": "L1D data cache lines in M state evicted due to replacement.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x51",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "L1D.ALL_M_REPLACEMENT",
"SampleAfterValue": "2000003",
"BriefDescription": "Cache lines in M state evicted out of L1D due to Snoop HitM or dirty line replacement.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding Demand Data Read transactions in uncore queue.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_C6",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
"CounterMask": "6",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_RFO",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding RFO store transactions in SuperQueue (SQ), queue to uncore.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when offcore outstanding cacheable Core Data Read transactions are present in SuperQueue (SQ), queue to uncore.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x63",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "LOCK_CYCLES.CACHE_LOCK_DURATION",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when L1D is locked.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD",
"SampleAfterValue": "100003",
"BriefDescription": "Demand Data Read requests sent to uncore.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD",
"SampleAfterValue": "100003",
"BriefDescription": "Cacheable and noncachaeble code read requests.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "OFFCORE_REQUESTS.DEMAND_RFO",
"SampleAfterValue": "100003",
"BriefDescription": "Demand RFO requests including regular RFOs, locks, ItoM.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB0",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OFFCORE_REQUESTS.ALL_DATA_RD",
"SampleAfterValue": "100003",
"BriefDescription": "Demand and prefetch data reads.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB2",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_BUFFER.SQ_FULL",
"SampleAfterValue": "2000003",
"BriefDescription": "Cases when offcore requests buffer cannot take more entries for core.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xBF",
"Counter": "0,1,2,3",
"UMask": "0x5",
"EventName": "L1D_BLOCKS.BANK_CONFLICT_CYCLES",
"SampleAfterValue": "100003",
"BriefDescription": "Cycles when dispatched loads are cancelled due to L1D bank conflicts with other load ports.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PEBS": "1",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x11",
"EventName": "MEM_UOPS_RETIRED.STLB_MISS_LOADS",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops that miss the STLB. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x12",
"EventName": "MEM_UOPS_RETIRED.STLB_MISS_STORES",
"SampleAfterValue": "100003",
"BriefDescription": "Retired store uops that miss the STLB. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x21",
"EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
"SampleAfterValue": "100007",
"BriefDescription": "Retired load uops with locked access. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K). (Precise Event - PEBS)",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x41",
"EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops that split across a cacheline boundary. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K). (Precise Event - PEBS)",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x42",
"EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
"SampleAfterValue": "100003",
"BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts the number of load uops retired (Precise Event)",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x81",
"EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
"SampleAfterValue": "2000003",
"BriefDescription": "All retired load uops. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts the number of store uops retired. (Precise Event - PEBS)",
"EventCode": "0xD0",
"Counter": "0,1,2,3",
"UMask": "0x82",
"EventName": "MEM_UOPS_RETIRED.ALL_STORES",
"SampleAfterValue": "2000003",
"BriefDescription": "All retired store uops. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
"SampleAfterValue": "2000003",
"BriefDescription": "Retired load uops with L1 cache hits as data sources. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops with L2 cache hits as data sources. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired load uops that hit in the last-level (L3) cache without snoops required. (Precise Event - PEBS)",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "MEM_LOAD_UOPS_RETIRED.LLC_HIT",
"SampleAfterValue": "50021",
"BriefDescription": "Retired load uops which data sources were data hits in LLC without snoops required. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD1",
"Counter": "0,1,2,3",
"UMask": "0x40",
"EventName": "MEM_LOAD_UOPS_RETIRED.HIT_LFB",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_MISS",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were LLC hit and cross-core snoop missed in on-pkg core cache. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired load uops that hit in the last-level cache (L3) and were found in a non-modified state in a neighboring core's private cache (same package). Since the last level cache is inclusive, hits to the L3 may require snooping the private L2 caches of any cores on the same socket that have the line. In this case, a snoop was required, and another L2 had the line in a non-modified state. (Precise Event - PEBS)",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HIT",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were LLC and cross-core snoop hits in on-pkg core cache. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired load uops that hit in the last-level cache (L3) and were found in a non-modified state in a neighboring core's private cache (same package). Since the last level cache is inclusive, hits to the L3 may require snooping the private L2 caches of any cores on the same socket that have the line. In this case, a snoop was required, and another L2 had the line in a modified state, so the line had to be invalidated in that L2 cache and transferred to the requesting L2. (Precise Event - PEBS)",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HITM",
"SampleAfterValue": "20011",
"BriefDescription": "Retired load uops which data sources were HitM responses from shared LLC. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"EventCode": "0xD2",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_NONE",
"SampleAfterValue": "100003",
"BriefDescription": "Retired load uops which data sources were hits in LLC without snoops required. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"PEBS": "1",
"PublicDescription": "This event counts retired demand loads that missed the last-level (L3) cache. This means that the load is usually satisfied from memory in a client system or possibly from the remote socket in a server. Demand loads are non speculative load uops. (Precise Event - PEBS)",
"EventCode": "0xD4",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MEM_LOAD_UOPS_MISC_RETIRED.LLC_MISS",
"SampleAfterValue": "100007",
"BriefDescription": "Retired load uops with unknown information as data source in cache serviced the load. (Precise Event - PEBS).",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xF0",
"Counter": "0,1,2,3",
@ -622,24 +727,6 @@
"BriefDescription": "Dirty L2 cache lines filling the L2.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x2E",
"Counter": "0,1,2,3",
"UMask": "0x41",
"EventName": "LONGEST_LAT_CACHE.MISS",
"SampleAfterValue": "100003",
"BriefDescription": "Core-originated cacheable demand requests missed LLC.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x2E",
"Counter": "0,1,2,3",
"UMask": "0x4f",
"EventName": "LONGEST_LAT_CACHE.REFERENCE",
"SampleAfterValue": "100003",
"BriefDescription": "Core-originated cacheable demand requests that refer to LLC.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xF4",
"Counter": "0,1,2,3",
@ -649,93 +736,6 @@
"BriefDescription": "Split locks in SQ.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x3",
"EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD",
"SampleAfterValue": "200003",
"BriefDescription": "Demand Data Read requests.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0xc",
"EventName": "L2_RQSTS.ALL_RFO",
"SampleAfterValue": "200003",
"BriefDescription": "RFO requests to L2 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0x30",
"EventName": "L2_RQSTS.ALL_CODE_RD",
"SampleAfterValue": "200003",
"BriefDescription": "L2 code requests.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x24",
"Counter": "0,1,2,3",
"UMask": "0xc0",
"EventName": "L2_RQSTS.ALL_PF",
"SampleAfterValue": "200003",
"BriefDescription": "Requests from L2 hardware prefetchers.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xBF",
"Counter": "0,1,2,3",
"UMask": "0x5",
"EventName": "L1D_BLOCKS.BANK_CONFLICT_CYCLES",
"SampleAfterValue": "100003",
"BriefDescription": "Cycles when dispatched loads are cancelled due to L1D bank conflicts with other load ports.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
"SampleAfterValue": "2000003",
"BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x60",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_C6",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
"CounterMask": "6",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x48",
"Counter": "2",
"UMask": "0x1",
"AnyThread": "1",
"EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
"CounterMask": "1",
"CounterHTOff": "2"
},
{
"EventCode": "0x48",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "L1D_PEND_MISS.FB_FULL",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x10003c0244",
@ -1825,7 +1825,7 @@
"EventName": "OFFCORE_RESPONSE.DATA_IN.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = DATA_INTO_CORE and RESPONSE = ANY_RESPONSE",
"BriefDescription": "REQUEST = DATA_INTO_CORE and RESPONSE = ANY_RESPONSE",
"CounterHTOff": "0,1,2,3"
},
{
@ -1837,7 +1837,7 @@
"EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT_M.HITM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = DEMAND_RFO and RESPONSE = LLC_HIT_M and SNOOP = HITM",
"BriefDescription": "REQUEST = DEMAND_RFO and RESPONSE = LLC_HIT_M and SNOOP = HITM",
"CounterHTOff": "0,1,2,3"
},
{
@ -1849,7 +1849,7 @@
"EventName": "OFFCORE_RESPONSE.PF_IFETCH.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_RFO and RESPONSE = ANY_RESPONSE",
"BriefDescription": "REQUEST = PF_RFO and RESPONSE = ANY_RESPONSE",
"CounterHTOff": "0,1,2,3"
},
{
@ -1861,7 +1861,7 @@
"EventName": "OFFCORE_RESPONSE.PF_L_DATA_RD.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_LLC_DATA_RD and RESPONSE = ANY_RESPONSE",
"BriefDescription": "REQUEST = PF_LLC_DATA_RD and RESPONSE = ANY_RESPONSE",
"CounterHTOff": "0,1,2,3"
},
{
@ -1873,7 +1873,7 @@
"EventName": "OFFCORE_RESPONSE.PF_L_IFETCH.ANY_RESPONSE",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_LLC_IFETCH and RESPONSE = ANY_RESPONSE",
"BriefDescription": "REQUEST = PF_LLC_IFETCH and RESPONSE = ANY_RESPONSE",
"CounterHTOff": "0,1,2,3"
}
]

126
tools/perf/pmu-events/arch/x86/sandybridge/floating-point.json
View file

@ -1,67 +1,4 @@
[
{
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OTHER_ASSISTS.AVX_STORE",
"SampleAfterValue": "100003",
"BriefDescription": "Number of GSSE memory assist for stores. GSSE microcode assist is being invoked whenever the hardware is unable to properly handle GSSE-256b operations.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "OTHER_ASSISTS.AVX_TO_SSE",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x20",
"EventName": "OTHER_ASSISTS.SSE_TO_AVX",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "FP_ASSIST.X87_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to output value.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "FP_ASSIST.X87_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to input value.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "FP_ASSIST.SIMD_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to Output values.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "FP_ASSIST.SIMD_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to input values.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x10",
"Counter": "0,1,2,3",
@ -125,6 +62,69 @@
"BriefDescription": "Number of AVX-256 Computational FP double precision uops issued this cycle.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "OTHER_ASSISTS.AVX_STORE",
"SampleAfterValue": "100003",
"BriefDescription": "Number of GSSE memory assist for stores. GSSE microcode assist is being invoked whenever the hardware is unable to properly handle GSSE-256b operations.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "OTHER_ASSISTS.AVX_TO_SSE",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from AVX-256 to legacy SSE when penalty applicable.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC1",
"Counter": "0,1,2,3",
"UMask": "0x20",
"EventName": "OTHER_ASSISTS.SSE_TO_AVX",
"SampleAfterValue": "100003",
"BriefDescription": "Number of transitions from SSE to AVX-256 when penalty applicable.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "FP_ASSIST.X87_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to output value.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "FP_ASSIST.X87_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of X87 assists due to input value.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "FP_ASSIST.SIMD_OUTPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to Output values.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "FP_ASSIST.SIMD_INPUT",
"SampleAfterValue": "100003",
"BriefDescription": "Number of SIMD FP assists due to input values.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCA",
"Counter": "0,1,2,3",

326
tools/perf/pmu-events/arch/x86/sandybridge/frontend.json
View file

@ -1,23 +1,4 @@
[
{
"EventCode": "0x80",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "ICACHE.HIT",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Reads. both cacheable and noncacheable, including UC fetches.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of instruction cache, streaming buffer and victim cache misses. Counting includes unchacheable accesses.",
"EventCode": "0x80",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "ICACHE.MISSES",
"SampleAfterValue": "200003",
"BriefDescription": "Instruction cache, streaming buffer and victim cache misses.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
@ -36,6 +17,16 @@
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "IDQ.MITE_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
@ -45,6 +36,16 @@
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "IDQ.DSB_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
@ -54,6 +55,47 @@
"BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "IDQ.MS_DSB_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EdgeDetect": "1",
"EventName": "IDQ.MS_DSB_OCCUR",
"SampleAfterValue": "2000003",
"BriefDescription": "Deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while Microcode Sequenser (MS) is busy.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x18",
"EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops.",
"CounterMask": "4",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x18",
"EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
@ -63,6 +105,26 @@
"BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x24",
"EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles MITE is delivering 4 Uops.",
"CounterMask": "4",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x24",
"EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles MITE is delivering any Uop.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
@ -73,7 +135,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts cycles during which the microcode sequencer assisted the front-end in delivering uops. Microcode assists are used for complex instructions or scenarios that can't be handled by the standard decoder. Using other instructions, if possible, will usually improve performance. See the Intel? 64 and IA-32 Architectures Optimization Reference Manual for more information.",
"PublicDescription": "This event counts cycles during which the microcode sequencer assisted the front-end in delivering uops. Microcode assists are used for complex instructions or scenarios that can't be handled by the standard decoder. Using other instructions, if possible, will usually improve performance. See the Intel\u00ae 64 and IA-32 Architectures Optimization Reference Manual for more information.",
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x30",
@ -83,6 +145,45 @@
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x30",
"EdgeDetect": "1",
"EventName": "IDQ.MS_SWITCHES",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x3c",
"EventName": "IDQ.MITE_ALL_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x80",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "ICACHE.HIT",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of Instruction Cache, Streaming Buffer and Victim Cache Reads. both cacheable and noncacheable, including UC fetches.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of instruction cache, streaming buffer and victim cache misses. Counting includes unchacheable accesses.",
"EventCode": "0x80",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "ICACHE.MISSES",
"SampleAfterValue": "200003",
"BriefDescription": "Instruction cache, streaming buffer and victim cache misses.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of uops not delivered to the back-end per cycle, per thread, when the back-end was not stalled. In the ideal case 4 uops can be delivered each cycle. The event counts the undelivered uops - so if 3 were delivered in one cycle, the counter would be incremented by 1 for that cycle (4 - 3). If the back-end is stalled, the count for this event is not incremented even when uops were not delivered, because the back-end would not have been able to accept them. This event is used in determining the front-end bound category of the top-down pipeline slots characterization.",
"EventCode": "0x9C",
@ -113,6 +214,48 @@
"CounterMask": "3",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x9C",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_2_UOP_DELIV.CORE",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with less than 2 uops delivered by the front end.",
"CounterMask": "2",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x9C",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with less than 3 uops delivered by the front end.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x9C",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_GE_1_UOP_DELIV.CORE",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when 1 or more uops were delivered to the by the front end.",
"CounterMask": "4",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x9C",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
"SampleAfterValue": "2000003",
"BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xAB",
"Counter": "0,1,2,3",
@ -150,118 +293,6 @@
"BriefDescription": "Cycles when Decode Stream Buffer (DSB) fill encounter more than 3 Decode Stream Buffer (DSB) lines.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "IDQ.MITE_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x8",
"EventName": "IDQ.DSB_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "IDQ.MS_DSB_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EdgeDetect": "1",
"EventName": "IDQ.MS_DSB_OCCUR",
"SampleAfterValue": "2000003",
"BriefDescription": "Deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while Microcode Sequenser (MS) is busy.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x9C",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_2_UOP_DELIV.CORE",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with less than 2 uops delivered by the front end.",
"CounterMask": "2",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x9C",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles with less than 3 uops delivered by the front end.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x9C",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_GE_1_UOP_DELIV.CORE",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when 1 or more uops were delivered to the by the front end.",
"CounterMask": "4",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x18",
"EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops.",
"CounterMask": "4",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x18",
"EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x24",
"EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles MITE is delivering 4 Uops.",
"CounterMask": "4",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x24",
"EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles MITE is delivering any Uop.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xAC",
"Counter": "0,1,2,3",
@ -270,36 +301,5 @@
"SampleAfterValue": "2000003",
"BriefDescription": "Cases of cancelling valid Decode Stream Buffer (DSB) fill not because of exceeding way limit.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x9C",
"Invert": "1",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
"SampleAfterValue": "2000003",
"BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x3c",
"EventName": "IDQ.MITE_ALL_UOPS",
"SampleAfterValue": "2000003",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"Counter": "0,1,2,3",
"UMask": "0x30",
"EdgeDetect": "1",
"EventName": "IDQ.MS_SWITCHES",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
"CounterMask": "1",
"CounterHTOff": "0,1,2,3,4,5,6,7"
}
]

68
tools/perf/pmu-events/arch/x86/sandybridge/memory.json
View file

@ -1,4 +1,31 @@
[
{
"EventCode": "0x05",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "MISALIGN_MEM_REF.LOADS",
"SampleAfterValue": "2000003",
"BriefDescription": "Speculative cache line split load uops dispatched to L1 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x05",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MISALIGN_MEM_REF.STORES",
"SampleAfterValue": "2000003",
"BriefDescription": "Speculative cache line split STA uops dispatched to L1 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xBE",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "PAGE_WALKS.LLC_MISS",
"SampleAfterValue": "100003",
"BriefDescription": "Number of any page walk that had a miss in LLC. Does not necessary cause a SUSPEND.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts the number of memory ordering Machine Clears detected. Memory Ordering Machine Clears can result from memory disambiguation, external snoops, or cross SMT-HW-thread snoop (stores) hitting load buffers. Machine clears can have a significant performance impact if they are happening frequently.",
"EventCode": "0xC3",
@ -125,33 +152,6 @@
"TakenAlone": "1",
"CounterHTOff": "3"
},
{
"EventCode": "0xBE",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "PAGE_WALKS.LLC_MISS",
"SampleAfterValue": "100003",
"BriefDescription": "Number of any page walk that had a miss in LLC. Does not necessary cause a SUSPEND.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x05",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "MISALIGN_MEM_REF.LOADS",
"SampleAfterValue": "2000003",
"BriefDescription": "Speculative cache line split load uops dispatched to L1 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x05",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "MISALIGN_MEM_REF.STORES",
"SampleAfterValue": "2000003",
"BriefDescription": "Speculative cache line split STA uops dispatched to L1 cache.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB7, 0xBB",
"MSRValue": "0x300400244",
@ -367,7 +367,7 @@
"EventName": "OFFCORE_RESPONSE.ANY_REQUEST.LLC_MISS_LOCAL.DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = ANY_REQUEST and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"BriefDescription": "REQUEST = ANY_REQUEST and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"CounterHTOff": "0,1,2,3"
},
{
@ -379,7 +379,7 @@
"EventName": "OFFCORE_RESPONSE.DATA_IN_SOCKET.LLC_MISS_LOCAL.ANY_LLC_HIT",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = DATA_IN_SOCKET and RESPONSE = LLC_MISS_LOCAL and SNOOP = ANY_LLC_HIT",
"BriefDescription": "REQUEST = DATA_IN_SOCKET and RESPONSE = LLC_MISS_LOCAL and SNOOP = ANY_LLC_HIT",
"CounterHTOff": "0,1,2,3"
},
{
@ -391,7 +391,7 @@
"EventName": "OFFCORE_RESPONSE.DEMAND_IFETCH.LLC_MISS_LOCAL.DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = DEMAND_IFETCH and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"BriefDescription": "REQUEST = DEMAND_IFETCH and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"CounterHTOff": "0,1,2,3"
},
{
@ -403,7 +403,7 @@
"EventName": "OFFCORE_RESPONSE.PF_DATA_RD.LLC_MISS_LOCAL.DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_DATA_RD and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"BriefDescription": "REQUEST = PF_DATA_RD and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"CounterHTOff": "0,1,2,3"
},
{
@ -415,7 +415,7 @@
"EventName": "OFFCORE_RESPONSE.PF_IFETCH.LLC_MISS_LOCAL.DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_RFO and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"BriefDescription": "REQUEST = PF_RFO and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"CounterHTOff": "0,1,2,3"
},
{
@ -427,7 +427,7 @@
"EventName": "OFFCORE_RESPONSE.PF_L_DATA_RD.LLC_MISS_LOCAL.DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_LLC_DATA_RD and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"BriefDescription": "REQUEST = PF_LLC_DATA_RD and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"CounterHTOff": "0,1,2,3"
},
{
@ -439,7 +439,7 @@
"EventName": "OFFCORE_RESPONSE.PF_L_IFETCH.LLC_MISS_LOCAL.DRAM",
"MSRIndex": "0x1a6,0x1a7",
"SampleAfterValue": "100003",
"BriefDescription": " REQUEST = PF_LLC_IFETCH and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"BriefDescription": "REQUEST = PF_LLC_IFETCH and RESPONSE = LLC_MISS_LOCAL and SNOOP = DRAM",
"CounterHTOff": "0,1,2,3"
}
]

18
tools/perf/pmu-events/arch/x86/sandybridge/other.json
View file

@ -8,6 +8,15 @@
"BriefDescription": "Valid instructions written to IQ per cycle.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x4E",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "HW_PRE_REQ.DL1_MISS",
"SampleAfterValue": "2000003",
"BriefDescription": "Hardware Prefetch requests that miss the L1D cache. This accounts for both L1 streamer and IP-based (IPP) HW prefetchers. A request is being counted each time it access the cache & miss it, including if a block is applicable or if hit the Fill Buffer for .",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x5C",
"Counter": "0,1,2,3",
@ -37,15 +46,6 @@
"BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x4E",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "HW_PRE_REQ.DL1_MISS",
"SampleAfterValue": "2000003",
"BriefDescription": "Hardware Prefetch requests that miss the L1D cache. This accounts for both L1 streamer and IP-based (IPP) HW prefetchers. A request is being counted each time it access the cache & miss it, including if a block is applicable or if hit the Fill Buffer for .",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x63",
"Counter": "0,1,2,3",

1628
tools/perf/pmu-events/arch/x86/sandybridge/pipeline.json
View file

File diff suppressed because it is too large Load diff

144
tools/perf/pmu-events/arch/x86/sandybridge/snb-metrics.json
View file

@ -1,140 +1,226 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4) )",
"MetricGroup": "Frontend",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 32 * ( ICACHE.HIT + ICACHE.MISSES ) / 4 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ) )",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_DISPATCHED.THREAD / (( cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@ / 2 ) if #SMT_on else cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_DISPATCHED.THREAD / (( cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@ / 2) if #SMT_on else cpu@UOPS_DISPATCHED.CORE\\,cmask\\=1@)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2 * FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4 * ( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8 * SIMD_FP_256.PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_COMP_OPS_EXE.SSE_SCALAR_SINGLE + FP_COMP_OPS_EXE.SSE_SCALAR_DOUBLE ) + 2* FP_COMP_OPS_EXE.SSE_PACKED_DOUBLE + 4*( FP_COMP_OPS_EXE.SSE_PACKED_SINGLE + SIMD_FP_256.PACKED_DOUBLE ) + 8* SIMD_FP_256.PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "64 * ( arb@event\\=0x81\\,umask\\=0x1@ + arb@event\\=0x84\\,umask\\=0x1@ ) / 1000000 / duration_time / 1000",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

110
tools/perf/pmu-events/arch/x86/sandybridge/virtual-memory.json
View file

@ -1,59 +1,4 @@
[
{
"EventCode": "0xAE",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "ITLB.ITLB_FLUSH",
"SampleAfterValue": "100007",
"BriefDescription": "Flushing of the Instruction TLB (ITLB) pages, includes 4k/2M/4M pages.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x4F",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "EPT.WALK_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycle count for an Extended Page table walk. The Extended Page Directory cache is used by Virtual Machine operating systems while the guest operating systems use the standard TLB caches.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "ITLB_MISSES.MISS_CAUSES_A_WALK",
"SampleAfterValue": "100003",
"BriefDescription": "Misses at all ITLB levels that cause page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "ITLB_MISSES.WALK_COMPLETED",
"SampleAfterValue": "100003",
"BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event count cycles when Page Miss Handler (PMH) is servicing page walks caused by ITLB misses.",
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "ITLB_MISSES.WALK_DURATION",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when PMH is busy with page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "ITLB_MISSES.STLB_HIT",
"SampleAfterValue": "100003",
"BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x08",
"Counter": "0,1,2,3",
@ -128,6 +73,61 @@
"BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x4F",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "EPT.WALK_CYCLES",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycle count for an Extended Page table walk. The Extended Page Directory cache is used by Virtual Machine operating systems while the guest operating systems use the standard TLB caches.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "ITLB_MISSES.MISS_CAUSES_A_WALK",
"SampleAfterValue": "100003",
"BriefDescription": "Misses at all ITLB levels that cause page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x2",
"EventName": "ITLB_MISSES.WALK_COMPLETED",
"SampleAfterValue": "100003",
"BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event count cycles when Page Miss Handler (PMH) is servicing page walks caused by ITLB misses.",
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x4",
"EventName": "ITLB_MISSES.WALK_DURATION",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles when PMH is busy with page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x85",
"Counter": "0,1,2,3",
"UMask": "0x10",
"EventName": "ITLB_MISSES.STLB_HIT",
"SampleAfterValue": "100003",
"BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xAE",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "ITLB.ITLB_FLUSH",
"SampleAfterValue": "100007",
"BriefDescription": "Flushing of the Instruction TLB (ITLB) pages, includes 4k/2M/4M pages.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xBD",
"Counter": "0,1,2,3",

2
tools/perf/pmu-events/arch/x86/silvermont/cache.json
View file

@ -36,7 +36,7 @@
"BriefDescription": "L2 cache request misses"
},
{
"PublicDescription": "Counts cycles that fetch is stalled due to an outstanding ICache miss. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes due to an ICache miss. Note: this event is not the same as the total number of cycles spent retrieving instruction cache lines from the memory hierarchy.\r\nCounts cycles that fetch is stalled due to any reason. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes. This will include cycles due to an ITLB miss, ICache miss and other events. \r\n",
"PublicDescription": "Counts cycles that fetch is stalled due to an outstanding ICache miss. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes due to an ICache miss. Note: this event is not the same as the total number of cycles spent retrieving instruction cache lines from the memory hierarchy.\r\nCounts cycles that fetch is stalled due to any reason. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes. This will include cycles due to an ITLB miss, ICache miss and other events.",
"EventCode": "0x86",
"Counter": "0,1",
"UMask": "0x4",

20
tools/perf/pmu-events/arch/x86/silvermont/other.json Normal file
View file

@ -0,0 +1,20 @@
[
{
"PublicDescription": "Counts cycles that fetch is stalled due to an outstanding ITLB miss. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes due to an ITLB miss. Note: this event is not the same as page walk cycles to retrieve an instruction translation.",
"EventCode": "0x86",
"Counter": "0,1",
"UMask": "0x2",
"EventName": "FETCH_STALL.ITLB_FILL_PENDING_CYCLES",
"SampleAfterValue": "200003",
"BriefDescription": "Cycles code-fetch stalled due to an outstanding ITLB miss."
},
{
"PublicDescription": "Counts cycles that fetch is stalled due to any reason. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes. This will include cycles due to an ITLB miss, ICache miss and other events.",
"EventCode": "0x86",
"Counter": "0,1",
"UMask": "0x3f",
"EventName": "FETCH_STALL.ALL",
"SampleAfterValue": "200003",
"BriefDescription": "Cycles code-fetch stalled due to any reason."
}
]

5
tools/perf/pmu-events/arch/x86/silvermont/pipeline.json
View file

@ -210,7 +210,7 @@
"UMask": "0x4",
"EventName": "NO_ALLOC_CYCLES.MISPREDICTS",
"SampleAfterValue": "200003",
"BriefDescription": "Counts the number of cycles when no uops are allocated and the alloc pipe is stalled waiting for a mispredicted jump to retire. After the misprediction is detected, the front end will start immediately but the allocate pipe stalls until the mispredicted "
"BriefDescription": "Counts the number of cycles when no uops are allocated and the alloc pipe is stalled waiting for a mispredicted jump to retire. After the misprediction is detected, the front end will start immediately but the allocate pipe stalls until the mispredicted"
},
{
"EventCode": "0xCA",
@ -275,7 +275,6 @@
},
{
"PublicDescription": "This event counts the number of instructions that retire. For instructions that consist of multiple micro-ops, this event counts exactly once, as the last micro-op of the instruction retires. The event continues counting while instructions retire, including during interrupt service routines caused by hardware interrupts, faults or traps. Background: Modern microprocessors employ extensive pipelining and speculative techniques. Since sometimes an instruction is started but never completed, the notion of \"retirement\" is introduced. A retired instruction is one that commits its states. Or stated differently, an instruction might be abandoned at some point. No instruction is truly finished until it retires. This counter measures the number of completed instructions. The fixed event is INST_RETIRED.ANY and the programmable event is INST_RETIRED.ANY_P.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -284,7 +283,6 @@
},
{
"PublicDescription": "Counts the number of core cycles while the core is not in a halt state. The core enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time. For this reason this event may have a changing ratio with regards to time. In systems with a constant core frequency, this event can give you a measurement of the elapsed time while the core was not in halt state by dividing the event count by the core frequency. This event is architecturally defined and is a designated fixed counter. CPU_CLK_UNHALTED.CORE and CPU_CLK_UNHALTED.CORE_P use the core frequency which may change from time to time. CPU_CLK_UNHALTE.REF_TSC and CPU_CLK_UNHALTED.REF are not affected by core frequency changes but counts as if the core is running at the maximum frequency all the time. The fixed events are CPU_CLK_UNHALTED.CORE and CPU_CLK_UNHALTED.REF_TSC and the programmable events are CPU_CLK_UNHALTED.CORE_P and CPU_CLK_UNHALTED.REF.",
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.CORE",
@ -293,7 +291,6 @@
},
{
"PublicDescription": "Counts the number of reference cycles while the core is not in a halt state. The core enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time. This event is not affected by core frequency changes but counts as if the core is running at the maximum frequency all the time. Divide this event count by core frequency to determine the elapsed time while the core was not in halt state. Divide this event count by core frequency to determine the elapsed time while the core was not in halt state. This event is architecturally defined and is a designated fixed counter. CPU_CLK_UNHALTED.CORE and CPU_CLK_UNHALTED.CORE_P use the core frequency which may change from time to time. CPU_CLK_UNHALTE.REF_TSC and CPU_CLK_UNHALTED.REF are not affected by core frequency changes but counts as if the core is running at the maximum frequency all the time. The fixed events are CPU_CLK_UNHALTED.CORE and CPU_CLK_UNHALTED.REF_TSC and the programmable events are CPU_CLK_UNHALTED.CORE_P and CPU_CLK_UNHALTED.REF.",
"EventCode": "0x00",
"Counter": "Fixed counter 3",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",

2157
tools/perf/pmu-events/arch/x86/skylake/cache.json
View file

File diff suppressed because it is too large Load diff

14
tools/perf/pmu-events/arch/x86/skylake/frontend.json
View file

@ -177,7 +177,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when: a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread. b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions). c. Instruction Decode Queue (IDQ) delivers four uops.",
"PublicDescription": "Counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding \u201c4 \u2013 x\u201d when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when: a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread. b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions). c. Instruction Decode Queue (IDQ) delivers four uops.",
"EventCode": "0x9C",
"Counter": "0,1,2,3",
"UMask": "0x1",
@ -242,7 +242,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. MM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.Penalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
"PublicDescription": "Counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. MM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.Penalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 0\u20132 cycles.",
"EventCode": "0xAB",
"Counter": "0,1,2,3",
"UMask": "0x2",
@ -253,7 +253,7 @@
},
{
"PEBS": "1",
"PublicDescription": "Counts retired Instructions that experienced DSB (Decode stream buffer i.e. the decoded instruction-cache) miss. \r\n",
"PublicDescription": "Counts retired Instructions that experienced DSB (Decode stream buffer i.e. the decoded instruction-cache) miss.",
"EventCode": "0xC6",
"MSRValue": "0x11",
"Counter": "0,1,2,3",
@ -360,7 +360,7 @@
},
{
"PEBS": "1",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 8 cycles. During this period the front-end delivered no uops. \r\n",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 8 cycles. During this period the front-end delivered no uops.",
"EventCode": "0xC6",
"MSRValue": "0x400806",
"Counter": "0,1,2,3",
@ -374,7 +374,7 @@
},
{
"PEBS": "1",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 16 cycles. During this period the front-end delivered no uops.\r\n",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 16 cycles. During this period the front-end delivered no uops.",
"EventCode": "0xC6",
"MSRValue": "0x401006",
"Counter": "0,1,2,3",
@ -388,7 +388,7 @@
},
{
"PEBS": "1",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 32 cycles. During this period the front-end delivered no uops.\r\n",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 32 cycles. During this period the front-end delivered no uops.",
"EventCode": "0xC6",
"MSRValue": "0x402006",
"Counter": "0,1,2,3",
@ -454,7 +454,7 @@
},
{
"PEBS": "1",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after the front-end had at least 1 bubble-slot for a period of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there was no RAT stall.\r\n",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after the front-end had at least 1 bubble-slot for a period of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there was no RAT stall.",
"EventCode": "0xC6",
"MSRValue": "0x100206",
"Counter": "0,1,2,3",

1121
tools/perf/pmu-events/arch/x86/skylake/memory.json
View file

File diff suppressed because it is too large Load diff

39
tools/perf/pmu-events/arch/x86/skylake/pipeline.json
View file

@ -1,7 +1,6 @@
[
{
"PublicDescription": "Counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, Counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. Notes: INST_RETIRED.ANY is counted by a designated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. INST_RETIRED.ANY_P is counted by a programmable counter and it is an architectural performance event. Counting: Faulting executions of GETSEC/VM entry/VM Exit/MWait will not count as retired instructions.",
"EventCode": "0x00",
"Counter": "Fixed counter 0",
"UMask": "0x1",
"EventName": "INST_RETIRED.ANY",
@ -11,7 +10,6 @@
},
{
"PublicDescription": "Counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"EventName": "CPU_CLK_UNHALTED.THREAD",
@ -20,7 +18,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"Counter": "Fixed counter 1",
"UMask": "0x2",
"AnyThread": "1",
@ -31,7 +28,6 @@
},
{
"PublicDescription": "Counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. Note: On all current platforms this event stops counting during 'throttling (TM)' states duty off periods the processor is 'halted'. The counter update is done at a lower clock rate then the core clock the overflow status bit for this counter may appear 'sticky'. After the counter has overflowed and software clears the overflow status bit and resets the counter to less than MAX. The reset value to the counter is not clocked immediately so the overflow status bit will flip 'high (1)' and generate another PMI (if enabled) after which the reset value gets clocked into the counter. Therefore, software will get the interrupt, read the overflow status bit '1 for bit 34 while the counter value is less than MAX. Software should ignore this case.",
"EventCode": "0x00",
"Counter": "Fixed counter 2",
"UMask": "0x3",
"EventName": "CPU_CLK_UNHALTED.REF_TSC",
@ -121,7 +117,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts the number of Blend Uops issued by the Resource Allocation Table (RAT) to the reservation station (RS) in order to preserve upper bits of vector registers. Starting with the Skylake microarchitecture, these Blend uops are needed since every Intel SSE instruction executed in Dirty Upper State needs to preserve bits 128-255 of the destination register. For more information, refer to Mixing Intel AVX and Intel SSE Code section of the Optimization Guide.",
"PublicDescription": "Counts the number of Blend Uops issued by the Resource Allocation Table (RAT) to the reservation station (RS) in order to preserve upper bits of vector registers. Starting with the Skylake microarchitecture, these Blend uops are needed since every Intel SSE instruction executed in Dirty Upper State needs to preserve bits 128-255 of the destination register. For more information, refer to \u201cMixing Intel AVX and Intel SSE Code\u201d section of the Optimization Guide.",
"EventCode": "0x0E",
"Counter": "0,1,2,3",
"UMask": "0x2",
@ -247,6 +243,16 @@
"BriefDescription": "Demand load dispatches that hit L1D fill buffer (FB) allocated for software prefetch.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This event counts cycles during which the microcode scoreboard stalls happen.",
"EventCode": "0x59",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "PARTIAL_RAT_STALLS.SCOREBOARD",
"SampleAfterValue": "2000003",
"BriefDescription": "Cycles where the pipeline is stalled due to serializing operations.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts cycles during which the reservation station (RS) is empty for the thread.; Note: In ST-mode, not active thread should drive 0. This is usually caused by severely costly branch mispredictions, or allocator/FE issues.",
"EventCode": "0x5E",
@ -361,8 +367,8 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts resource-related stall cycles. Reasons for stalls can be as follows:a. *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots).b. *any* u-arch structure got empty (like INT/SIMD FreeLists).c. FPU control word (FPCW), MXCSR.and others. This counts cycles that the pipeline back-end blocked uop delivery from the front-end.",
"EventCode": "0xA2",
"PublicDescription": "Counts resource-related stall cycles.",
"EventCode": "0xa2",
"Counter": "0,1,2,3",
"UMask": "0x1",
"EventName": "RESOURCE_STALLS.ANY",
@ -735,7 +741,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts cycles without actually retired uops.",
"PublicDescription": "This event counts cycles without actually retired uops.",
"EventCode": "0xC2",
"Invert": "1",
"Counter": "0,1,2,3",
@ -759,6 +765,7 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Number of machine clears (nukes) of any type.",
"EventCode": "0xC3",
"Counter": "0,1,2,3",
"UMask": "0x1",
@ -839,14 +846,15 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts not taken branch instructions retired.",
"PEBS": "1",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts not taken branch instructions retired.",
"EventCode": "0xC4",
"Counter": "0,1,2,3",
"UMask": "0x10",
"Errata": "SKL091",
"EventName": "BR_INST_RETIRED.NOT_TAKEN",
"SampleAfterValue": "400009",
"BriefDescription": "Not taken branch instructions retired.",
"BriefDescription": "Counts all not taken macro branch instructions retired.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -924,7 +932,7 @@
"UMask": "0x20",
"EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
"SampleAfterValue": "400009",
"BriefDescription": "Number of near branch instructions retired that were mispredicted and taken. ",
"BriefDescription": "Number of near branch instructions retired that were mispredicted and taken.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
@ -937,6 +945,15 @@
"BriefDescription": "Increments whenever there is an update to the LBR array.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCC",
"Counter": "0,1,2,3",
"UMask": "0x40",
"EventName": "ROB_MISC_EVENTS.PAUSE_INST",
"SampleAfterValue": "2000003",
"BriefDescription": "Number of retired PAUSE instructions (that do not end up with a VMExit to the VMM; TSX aborted Instructions may be counted). This event is not supported on first SKL and KBL products.",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"PublicDescription": "Counts the number of times the front-end is resteered when it finds a branch instruction in a fetch line. This occurs for the first time a branch instruction is fetched or when the branch is not tracked by the BPU (Branch Prediction Unit) anymore.",
"EventCode": "0xE6",

274
tools/perf/pmu-events/arch/x86/skylake/skl-metrics.json
View file

@ -1,164 +1,364 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ((UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 64 * ( ICACHE_64B.IFTAG_HIT + ICACHE_64B.IFTAG_MISS ) / 4.1) )",
"MetricGroup": "Frontend",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 64 * ( ICACHE_64B.IFTAG_HIT + ICACHE_64B.IFTAG_MISS ) / 4.1 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ))",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_INST_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_INST_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( UOPS_EXECUTED.CORE_CYCLES_GE_1 / 2 ) if #SMT_on else UOPS_EXECUTED.CORE_CYCLES_GE_1)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( UOPS_EXECUTED.CORE_CYCLES_GE_1 / 2) if #SMT_on else UOPS_EXECUTED.CORE_CYCLES_GE_1)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE_16B.IFDATA_STALL - ICACHE_64B.IFTAG_STALL ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) * (( INT_MISC.CLEAR_RESTEER_CYCLES + 9 * BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) ) * (4 * cycles) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "Branch_Misprediction_Cost"
},
{
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) * (( INT_MISC.CLEAR_RESTEER_CYCLES + 9 * BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) ) * (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts_SMT",
"MetricName": "Branch_Misprediction_Cost_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_RETIRED.L1_MISS + MEM_LOAD_RETIRED.FB_HIT )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( L1D_PEND_MISS.PENDING_CYCLES_ANY / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * cycles )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles) )",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )) )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "64 * OFFCORE_REQUESTS.ALL_REQUESTS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Access_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * L2_RQSTS.MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * ( L2_RQSTS.REFERENCES - L2_RQSTS.MISS ) / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2* FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4*( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8* FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "64 * ( arb@event\\=0x81\\,umask\\=0x1@ + arb@event\\=0x84\\,umask\\=0x1@ ) / 1000000 / duration_time / 1000",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "arb@event\\=0x80\\,umask\\=0x2@ / arb@event\\=0x80\\,umask\\=0x2\\,thresh\\=1@",
"BriefDescription": "Average number of parallel data read requests to external memory. Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_Parallel_Reads"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

930
tools/perf/pmu-events/arch/x86/skylakex/cache.json
View file

File diff suppressed because it is too large Load diff

2
tools/perf/pmu-events/arch/x86/skylakex/floating-point.json
View file

@ -59,7 +59,6 @@
"BriefDescription": "Number of Packed Double-Precision FP arithmetic instructions (Use operation multiplier of 8)",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE",
"PublicDescription": "Number of Packed Double-Precision FP arithmetic instructions (Use operation multiplier of 8).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -69,7 +68,6 @@
"BriefDescription": "Number of Packed Single-Precision FP arithmetic instructions (Use operation multiplier of 16)",
"Counter": "0,1,2,3",
"EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE",
"PublicDescription": "Number of Packed Single-Precision FP arithmetic instructions (Use operation multiplier of 16).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},

358
tools/perf/pmu-events/arch/x86/skylakex/frontend.json
View file

@ -1,14 +1,4 @@
[
{
"EventCode": "0x79",
"UMask": "0x4",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path",
"Counter": "0,1,2,3",
"EventName": "IDQ.MITE_UOPS",
"PublicDescription": "Counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may 'bypass' the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x4",
@ -22,11 +12,11 @@
},
{
"EventCode": "0x79",
"UMask": "0x8",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
"UMask": "0x4",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path",
"Counter": "0,1,2,3",
"EventName": "IDQ.DSB_UOPS",
"PublicDescription": "Counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may 'bypass' the IDQ.",
"EventName": "IDQ.MITE_UOPS",
"PublicDescription": "Counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may 'bypass' the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -41,6 +31,16 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x8",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
"Counter": "0,1,2,3",
"EventName": "IDQ.DSB_UOPS",
"PublicDescription": "Counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may 'bypass' the IDQ.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x10",
@ -55,22 +55,22 @@
{
"EventCode": "0x79",
"UMask": "0x18",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop",
"Counter": "0,1,2,3",
"EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
"CounterMask": "4",
"PublicDescription": "Counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Count includes uops that may 'bypass' the IDQ.",
"EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
"CounterMask": "1",
"PublicDescription": "Counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Count includes uops that may 'bypass' the IDQ.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x18",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering any Uop",
"BriefDescription": "Cycles Decode Stream Buffer (DSB) is delivering 4 Uops",
"Counter": "0,1,2,3",
"EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
"CounterMask": "1",
"PublicDescription": "Counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Count includes uops that may 'bypass' the IDQ.",
"EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
"CounterMask": "4",
"PublicDescription": "Counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Count includes uops that may 'bypass' the IDQ.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -87,22 +87,22 @@
{
"EventCode": "0x79",
"UMask": "0x24",
"BriefDescription": "Cycles MITE is delivering 4 Uops",
"BriefDescription": "Cycles MITE is delivering any Uop",
"Counter": "0,1,2,3",
"EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
"CounterMask": "4",
"PublicDescription": "Counts the number of cycles 4 uops were delivered to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipeline) path. Counting includes uops that may 'bypass' the IDQ. During these cycles uops are not being delivered from the Decode Stream Buffer (DSB).",
"EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
"CounterMask": "1",
"PublicDescription": "Counts the number of cycles uops were delivered to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipeline) path. Counting includes uops that may 'bypass' the IDQ. During these cycles uops are not being delivered from the Decode Stream Buffer (DSB).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x24",
"BriefDescription": "Cycles MITE is delivering any Uop",
"BriefDescription": "Cycles MITE is delivering 4 Uops",
"Counter": "0,1,2,3",
"EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
"CounterMask": "1",
"PublicDescription": "Counts the number of cycles uops were delivered to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipeline) path. Counting includes uops that may 'bypass' the IDQ. During these cycles uops are not being delivered from the Decode Stream Buffer (DSB).",
"EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
"CounterMask": "4",
"PublicDescription": "Counts the number of cycles 4 uops were delivered to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipeline) path. Counting includes uops that may 'bypass' the IDQ. During these cycles uops are not being delivered from the Decode Stream Buffer (DSB).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -117,6 +117,16 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x30",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
"Counter": "0,1,2,3",
"EventName": "IDQ.MS_UOPS",
"PublicDescription": "Counts the total number of uops delivered by the Microcode Sequencer (MS). Any instruction over 4 uops will be delivered by the MS. Some instructions such as transcendentals may additionally generate uops from the MS.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EdgeDetect": "1",
"EventCode": "0x79",
@ -129,16 +139,6 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x79",
"UMask": "0x30",
"BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
"Counter": "0,1,2,3",
"EventName": "IDQ.MS_UOPS",
"PublicDescription": "Counts the total number of uops delivered by the Microcode Sequencer (MS). Any instruction over 4 uops will be delivered by the MS. Some instructions such as transcendentals may additionally generate uops from the MS.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x80",
"UMask": "0x4",
@ -177,34 +177,24 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0x9C",
"UMask": "0x1",
"BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled",
"BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.",
"Counter": "0,1,2,3",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CORE",
"PublicDescription": "Counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when: a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread. b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions). c. Instruction Decode Queue (IDQ) delivers four uops.",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
"CounterMask": "1",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x9C",
"UMask": "0x1",
"BriefDescription": "Cycles per thread when 4 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled",
"BriefDescription": "Cycles with less than 3 uops delivered by the front end.",
"Counter": "0,1,2,3",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE",
"CounterMask": "4",
"PublicDescription": "Counts, on the per-thread basis, cycles when no uops are delivered to Resource Allocation Table (RAT). IDQ_Uops_Not_Delivered.core =4.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x9C",
"UMask": "0x1",
"BriefDescription": "Cycles per thread when 3 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled",
"Counter": "0,1,2,3",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_1_UOP_DELIV.CORE",
"CounterMask": "3",
"PublicDescription": "Counts, on the per-thread basis, cycles when less than 1 uop is delivered to Resource Allocation Table (RAT). IDQ_Uops_Not_Delivered.core >= 3.",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
"CounterMask": "1",
"PublicDescription": "Cycles with less than 3 uops delivered by the front-end.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -222,22 +212,32 @@
{
"EventCode": "0x9C",
"UMask": "0x1",
"BriefDescription": "Cycles with less than 3 uops delivered by the front end.",
"BriefDescription": "Cycles per thread when 3 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled",
"Counter": "0,1,2,3",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_3_UOP_DELIV.CORE",
"CounterMask": "1",
"PublicDescription": "Cycles with less than 3 uops delivered by the front-end.",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_LE_1_UOP_DELIV.CORE",
"CounterMask": "3",
"PublicDescription": "Counts, on the per-thread basis, cycles when less than 1 uop is delivered to Resource Allocation Table (RAT). IDQ_Uops_Not_Delivered.core >= 3.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0x9C",
"UMask": "0x1",
"BriefDescription": "Counts cycles FE delivered 4 uops or Resource Allocation Table (RAT) was stalling FE.",
"BriefDescription": "Cycles per thread when 4 or more uops are not delivered to Resource Allocation Table (RAT) when backend of the machine is not stalled",
"Counter": "0,1,2,3",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK",
"CounterMask": "1",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE",
"CounterMask": "4",
"PublicDescription": "Counts, on the per-thread basis, cycles when no uops are delivered to Resource Allocation Table (RAT). IDQ_Uops_Not_Delivered.core =4.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x9C",
"UMask": "0x1",
"BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled",
"Counter": "0,1,2,3",
"EventName": "IDQ_UOPS_NOT_DELIVERED.CORE",
"PublicDescription": "Counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding \u201c4 \u2013 x\u201d when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when: a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread. b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions). c. Instruction Decode Queue (IDQ) delivers four uops.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -247,86 +247,18 @@
"BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.",
"Counter": "0,1,2,3",
"EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES",
"PublicDescription": "Counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. MM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.Penalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
"PublicDescription": "Counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. MM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.Penalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 0\u20132 cycles.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired Instructions who experienced decode stream buffer (DSB - the decoded instruction-cache) miss. Precise Event.",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 4 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x11",
"MSRValue": "0x400406",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.DSB_MISS",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired Instructions that experienced DSB (Decode stream buffer i.e. the decoded instruction-cache) miss. \r\n",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired Instructions who experienced Instruction L1 Cache true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x12",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.L1I_MISS",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired Instructions who experienced Instruction L2 Cache true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x13",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.L2_MISS",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired Instructions who experienced iTLB true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x14",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.ITLB_MISS",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired Instructions that experienced iTLB (Instruction TLB) true miss.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired Instructions who experienced STLB (2nd level TLB) true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x15",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.STLB_MISS",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired Instructions that experienced STLB (2nd level TLB) true miss.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 2 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x400206",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_2",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_4",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
@ -348,11 +280,11 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 4 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 2 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x400406",
"MSRValue": "0x400206",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_4",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_2",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
@ -361,13 +293,13 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 8 cycles which was not interrupted by a back-end stall.",
"BriefDescription": "Retired Instructions who experienced STLB (2nd level TLB) true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x400806",
"MSRValue": "0x15",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_8",
"EventName": "FRONTEND_RETIRED.STLB_MISS",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 8 cycles. During this period the front-end delivered no uops. \r\n",
"PublicDescription": "Counts retired Instructions that experienced STLB (2nd level TLB) true miss.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
@ -375,13 +307,13 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 16 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired Instructions who experienced iTLB true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x401006",
"MSRValue": "0x14",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_16",
"EventName": "FRONTEND_RETIRED.ITLB_MISS",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 16 cycles. During this period the front-end delivered no uops.\r\n",
"PublicDescription": "Counts retired Instructions that experienced iTLB (Instruction TLB) true miss.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
@ -389,25 +321,11 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 32 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired Instructions who experienced Instruction L2 Cache true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x402006",
"MSRValue": "0x13",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_32",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 32 cycles. During this period the front-end delivered no uops.\r\n",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 64 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x404006",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_64",
"EventName": "FRONTEND_RETIRED.L2_MISS",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
@ -416,11 +334,65 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 128 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired Instructions who experienced Instruction L1 Cache true miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x408006",
"MSRValue": "0x12",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_128",
"EventName": "FRONTEND_RETIRED.L1I_MISS",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired Instructions who experienced decode stream buffer (DSB - the decoded instruction-cache) miss. Precise Event.",
"PEBS": "1",
"MSRValue": "0x11",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.DSB_MISS",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired Instructions that experienced DSB (Decode stream buffer i.e. the decoded instruction-cache) miss.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end had at least 3 bubble-slots for a period of 2 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x300206",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_3",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end had at least 1 bubble-slot for a period of 2 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x100206",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_1",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after the front-end had at least 1 bubble-slot for a period of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there was no RAT stall.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 512 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x420006",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_512",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
@ -442,11 +414,11 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 512 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 128 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x420006",
"MSRValue": "0x408006",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_512",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_128",
"MSRIndex": "0x3F7",
"TakenAlone": "1",
"SampleAfterValue": "100007",
@ -455,13 +427,12 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end had at least 1 bubble-slot for a period of 2 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 64 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x100206",
"MSRValue": "0x404006",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_1",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_64",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after the front-end had at least 1 bubble-slot for a period of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there was no RAT stall.\r\n",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
@ -469,12 +440,41 @@
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end had at least 3 bubble-slots for a period of 2 cycles which was not interrupted by a back-end stall. Precise Event.",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 32 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x300206",
"MSRValue": "0x402006",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_32",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 32 cycles. During this period the front-end delivered no uops.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 16 cycles which was not interrupted by a back-end stall. Precise Event.",
"PEBS": "1",
"MSRValue": "0x401006",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_16",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 16 cycles. During this period the front-end delivered no uops.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"
},
{
"EventCode": "0xC6",
"UMask": "0x1",
"BriefDescription": "Retired instructions that are fetched after an interval where the front-end delivered no uops for a period of 8 cycles which was not interrupted by a back-end stall.",
"PEBS": "1",
"MSRValue": "0x400806",
"Counter": "0,1,2,3",
"EventName": "FRONTEND_RETIRED.LATENCY_GE_8",
"MSRIndex": "0x3F7",
"PublicDescription": "Counts retired instructions that are delivered to the back-end after a front-end stall of at least 8 cycles. During this period the front-end delivered no uops.",
"TakenAlone": "1",
"SampleAfterValue": "100007",
"CounterHTOff": "0,1,2,3"

835
tools/perf/pmu-events/arch/x86/skylakex/memory.json
View file

File diff suppressed because it is too large Load diff

243
tools/perf/pmu-events/arch/x86/skylakex/pipeline.json
View file

@ -1,6 +1,5 @@
[
{
"EventCode": "0x00",
"UMask": "0x1",
"BriefDescription": "Instructions retired from execution.",
"Counter": "Fixed counter 0",
@ -10,7 +9,6 @@
"CounterHTOff": "Fixed counter 0"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when the thread is not in halt state",
"Counter": "Fixed counter 1",
@ -20,7 +18,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x2",
"BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
"Counter": "Fixed counter 1",
@ -30,7 +27,6 @@
"CounterHTOff": "Fixed counter 1"
},
{
"EventCode": "0x00",
"UMask": "0x3",
"BriefDescription": "Reference cycles when the core is not in halt state.",
"Counter": "Fixed counter 2",
@ -98,16 +94,6 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x0E",
"UMask": "0x1",
"BriefDescription": "Uops that Resource Allocation Table (RAT) issues to Reservation Station (RS)",
"Counter": "0,1,2,3",
"EventName": "UOPS_ISSUED.ANY",
"PublicDescription": "Counts the number of uops that the Resource Allocation Table (RAT) issues to the Reservation Station (RS).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0x0E",
@ -120,13 +106,23 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x0E",
"UMask": "0x1",
"BriefDescription": "Uops that Resource Allocation Table (RAT) issues to Reservation Station (RS)",
"Counter": "0,1,2,3",
"EventName": "UOPS_ISSUED.ANY",
"PublicDescription": "Counts the number of uops that the Resource Allocation Table (RAT) issues to the Reservation Station (RS).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x0E",
"UMask": "0x2",
"BriefDescription": "Uops inserted at issue-stage in order to preserve upper bits of vector registers.",
"Counter": "0,1,2,3",
"EventName": "UOPS_ISSUED.VECTOR_WIDTH_MISMATCH",
"PublicDescription": "Counts the number of Blend Uops issued by the Resource Allocation Table (RAT) to the reservation station (RS) in order to preserve upper bits of vector registers. Starting with the Skylake microarchitecture, these Blend uops are needed since every Intel SSE instruction executed in Dirty Upper State needs to preserve bits 128-255 of the destination register. For more information, refer to Mixing Intel AVX and Intel SSE Code section of the Optimization Guide.",
"PublicDescription": "Counts the number of Blend Uops issued by the Resource Allocation Table (RAT) to the reservation station (RS) in order to preserve upper bits of vector registers. Starting with the Skylake microarchitecture, these Blend uops are needed since every Intel SSE instruction executed in Dirty Upper State needs to preserve bits 128-255 of the destination register. For more information, refer to \u201cMixing Intel AVX and Intel SSE Code\u201d section of the Optimization Guide.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -203,19 +199,19 @@
{
"EventCode": "0x3C",
"UMask": "0x1",
"BriefDescription": "Core crystal clock cycles when the thread is unhalted.",
"BriefDescription": "Core crystal clock cycles when at least one thread on the physical core is unhalted.",
"Counter": "0,1,2,3",
"EventName": "CPU_CLK_UNHALTED.REF_XCLK",
"EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
"AnyThread": "1",
"SampleAfterValue": "2503",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x3C",
"UMask": "0x1",
"BriefDescription": "Core crystal clock cycles when at least one thread on the physical core is unhalted.",
"BriefDescription": "Core crystal clock cycles when the thread is unhalted.",
"Counter": "0,1,2,3",
"EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
"AnyThread": "1",
"EventName": "CPU_CLK_UNHALTED.REF_XCLK",
"SampleAfterValue": "2503",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -248,12 +244,12 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x5E",
"EventCode": "0x59",
"UMask": "0x1",
"BriefDescription": "Cycles when Reservation Station (RS) is empty for the thread",
"BriefDescription": "Cycles where the pipeline is stalled due to serializing operations.",
"Counter": "0,1,2,3",
"EventName": "RS_EVENTS.EMPTY_CYCLES",
"PublicDescription": "Counts cycles during which the reservation station (RS) is empty for the thread.; Note: In ST-mode, not active thread should drive 0. This is usually caused by severely costly branch mispredictions, or allocator/FE issues.",
"EventName": "PARTIAL_RAT_STALLS.SCOREBOARD",
"PublicDescription": "This event counts cycles during which the microcode scoreboard stalls happen.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -270,6 +266,16 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x5E",
"UMask": "0x1",
"BriefDescription": "Cycles when Reservation Station (RS) is empty for the thread",
"Counter": "0,1,2,3",
"EventName": "RS_EVENTS.EMPTY_CYCLES",
"PublicDescription": "Counts cycles during which the reservation station (RS) is empty for the thread.; Note: In ST-mode, not active thread should drive 0. This is usually caused by severely costly branch mispredictions, or allocator/FE issues.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0x87",
"UMask": "0x1",
@ -361,12 +367,12 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xA2",
"EventCode": "0xa2",
"UMask": "0x1",
"BriefDescription": "Resource-related stall cycles",
"Counter": "0,1,2,3",
"EventName": "RESOURCE_STALLS.ANY",
"PublicDescription": "Counts resource-related stall cycles. Reasons for stalls can be as follows:a. *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots).b. *any* u-arch structure got empty (like INT/SIMD FreeLists).c. FPU control word (FPCW), MXCSR.and others. This counts cycles that the pipeline back-end blocked uop delivery from the front-end.",
"PublicDescription": "Counts resource-related stall cycles.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -519,17 +525,6 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xA8",
"UMask": "0x1",
"BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
"Counter": "0,1,2,3",
"EventName": "LSD.CYCLES_ACTIVE",
"CounterMask": "1",
"PublicDescription": "Counts the cycles when at least one uop is delivered by the LSD (Loop-stream detector).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xA8",
"UMask": "0x1",
@ -541,13 +536,58 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xA8",
"UMask": "0x1",
"BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
"Counter": "0,1,2,3",
"EventName": "LSD.CYCLES_ACTIVE",
"CounterMask": "1",
"PublicDescription": "Counts the cycles when at least one uop is delivered by the LSD (Loop-stream detector).",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
"BriefDescription": "Cycles where at least 4 uops were executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.THREAD",
"PublicDescription": "Number of uops to be executed per-thread each cycle.",
"EventName": "UOPS_EXECUTED.CYCLES_GE_4_UOPS_EXEC",
"CounterMask": "4",
"PublicDescription": "Cycles where at least 4 uops were executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 3 uops were executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC",
"CounterMask": "3",
"PublicDescription": "Cycles where at least 3 uops were executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 2 uops were executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC",
"CounterMask": "2",
"PublicDescription": "Cycles where at least 2 uops were executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 1 uop was executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC",
"CounterMask": "1",
"PublicDescription": "Cycles where at least 1 uop was executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -566,44 +606,10 @@
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 1 uop was executed per-thread",
"BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC",
"CounterMask": "1",
"PublicDescription": "Cycles where at least 1 uop was executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 2 uops were executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC",
"CounterMask": "2",
"PublicDescription": "Cycles where at least 2 uops were executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 3 uops were executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC",
"CounterMask": "3",
"PublicDescription": "Cycles where at least 3 uops were executed per-thread.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xB1",
"UMask": "0x1",
"BriefDescription": "Cycles where at least 4 uops were executed per-thread",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CYCLES_GE_4_UOPS_EXEC",
"CounterMask": "4",
"PublicDescription": "Cycles where at least 4 uops were executed per-thread.",
"EventName": "UOPS_EXECUTED.THREAD",
"PublicDescription": "Number of uops to be executed per-thread each cycle.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -618,11 +624,12 @@
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0xB1",
"UMask": "0x2",
"BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
"BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
"CounterMask": "1",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
@ -630,10 +637,10 @@
{
"EventCode": "0xB1",
"UMask": "0x2",
"BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
"BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
"CounterMask": "2",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
"CounterMask": "4",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -650,20 +657,19 @@
{
"EventCode": "0xB1",
"UMask": "0x2",
"BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
"BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
"CounterMask": "4",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
"CounterMask": "2",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0xB1",
"UMask": "0x2",
"BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
"BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
"Counter": "0,1,2,3",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
"EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
"CounterMask": "1",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
@ -724,28 +730,6 @@
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC2",
"UMask": "0x2",
"BriefDescription": "Retirement slots used.",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.RETIRE_SLOTS",
"PublicDescription": "Counts the retirement slots used.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0xC2",
"UMask": "0x2",
"BriefDescription": "Cycles without actually retired uops.",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.STALL_CYCLES",
"CounterMask": "1",
"PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts cycles without actually retired uops.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0xC2",
@ -758,6 +742,28 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"Invert": "1",
"EventCode": "0xC2",
"UMask": "0x2",
"BriefDescription": "Cycles without actually retired uops.",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.STALL_CYCLES",
"CounterMask": "1",
"PublicDescription": "This event counts cycles without actually retired uops.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xC2",
"UMask": "0x2",
"BriefDescription": "Retirement slots used.",
"Counter": "0,1,2,3",
"EventName": "UOPS_RETIRED.RETIRE_SLOTS",
"PublicDescription": "Counts the retirement slots used.",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EdgeDetect": "1",
"EventCode": "0xC3",
@ -766,6 +772,7 @@
"Counter": "0,1,2,3",
"EventName": "MACHINE_CLEARS.COUNT",
"CounterMask": "1",
"PublicDescription": "Number of machine clears (nukes) of any type.",
"SampleAfterValue": "100003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -841,11 +848,12 @@
{
"EventCode": "0xC4",
"UMask": "0x10",
"BriefDescription": "Not taken branch instructions retired.",
"BriefDescription": "Counts all not taken macro branch instructions retired.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_INST_RETIRED.NOT_TAKEN",
"Errata": "SKL091",
"PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts not taken branch instructions retired.",
"PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts not taken branch instructions retired.",
"SampleAfterValue": "400009",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
@ -919,7 +927,7 @@
{
"EventCode": "0xC5",
"UMask": "0x20",
"BriefDescription": "Number of near branch instructions retired that were mispredicted and taken. ",
"BriefDescription": "Number of near branch instructions retired that were mispredicted and taken.",
"PEBS": "1",
"Counter": "0,1,2,3",
"EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
@ -937,6 +945,15 @@
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xCC",
"UMask": "0x40",
"BriefDescription": "Number of retired PAUSE instructions (that do not end up with a VMExit to the VMM; TSX aborted Instructions may be counted). This event is not supported on first SKL and KBL products.",
"Counter": "0,1,2,3",
"EventName": "ROB_MISC_EVENTS.PAUSE_INST",
"SampleAfterValue": "2000003",
"CounterHTOff": "0,1,2,3,4,5,6,7"
},
{
"EventCode": "0xE6",
"UMask": "0x1",

304
tools/perf/pmu-events/arch/x86/skylakex/skx-metrics.json
View file

@ -1,164 +1,394 @@
[
{
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Frontend_Bound"
},
{
"MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. Frontend denotes the first part of the processor core responsible to fetch operations that are executed later on by the Backend part. Within the Frontend; a branch predictor predicts the next address to fetch; cache-lines are fetched from the memory subsystem; parsed into instructions; and lastly decoded into micro-ops (uops). Ideally the Frontend can issue 4 uops every cycle to the Backend. Frontend Bound denotes unutilized issue-slots when there is no Backend stall; i.e. bubbles where Frontend delivered no uops while Backend could have accepted them. For example; stalls due to instruction-cache misses would be categorized under Frontend Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where the processor's Frontend undersupplies its Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Frontend_Bound_SMT"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations",
"MetricGroup": "TopdownL1",
"MetricName": "Bad_Speculation"
},
{
"MetricExpr": "( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots wasted due to incorrect speculations. This include slots used to issue uops that do not eventually get retired and slots for which the issue-pipeline was blocked due to recovery from earlier incorrect speculation. For example; wasted work due to miss-predicted branches are categorized under Bad Speculation category. Incorrect data speculation followed by Memory Ordering Nukes is another example. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots wasted due to incorrect speculations. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Bad_Speculation_SMT"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * cycles)) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles)) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend",
"MetricGroup": "TopdownL1",
"MetricName": "Backend_Bound"
},
{
"MetricExpr": "1 - ( (IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) + (UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) )",
"PublicDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. Backend is the portion of the processor core where the out-of-order scheduler dispatches ready uops into their respective execution units; and once completed these uops get retired according to program order. For example; stalls due to data-cache misses or stalls due to the divider unit being overloaded are both categorized under Backend Bound. Backend Bound is further divided into two main categories: Memory Bound and Core Bound. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots where no uops are being delivered due to a lack of required resources for accepting new uops in the Backend. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Backend_Bound_SMT"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * cycles)",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. ",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired",
"MetricGroup": "TopdownL1",
"MetricName": "Retiring"
},
{
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))",
"PublicDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. Ideally; all pipeline slots would be attributed to the Retiring category. Retiring of 100% would indicate the maximum 4 uops retired per cycle has been achieved. Maximizing Retiring typically increases the Instruction-Per-Cycle metric. Note that a high Retiring value does not necessary mean there is no room for more performance. For example; Microcode assists are categorized under Retiring. They hurt performance and can often be avoided. SMT version; use when SMT is enabled and measuring per logical CPU.",
"BriefDescription": "This category represents fraction of slots utilized by useful work i.e. issued uops that eventually get retired. SMT version; use when SMT is enabled and measuring per logical CPU.",
"MetricGroup": "TopdownL1_SMT",
"MetricName": "Retiring_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Instructions Per Cycle (per logical thread)",
"MetricGroup": "TopDownL1",
"MetricName": "IPC"
},
{
"BriefDescription": "Uops Per Instruction",
"MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY",
"MetricGroup": "Pipeline",
"BriefDescription": "Uops Per Instruction",
"MetricGroup": "Pipeline;Retiring",
"MetricName": "UPI"
},
{
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely consumed by program instructions",
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ((UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 64 * ( ICACHE_64B.IFTAG_HIT + ICACHE_64B.IFTAG_MISS ) / 4.1) )",
"MetricGroup": "Frontend",
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Instruction per taken branch",
"MetricGroup": "Branches;PGO",
"MetricName": "IpTB"
},
{
"MetricExpr": "BR_INST_RETIRED.ALL_BRANCHES / BR_INST_RETIRED.NEAR_TAKEN",
"BriefDescription": "Branch instructions per taken branch. ",
"MetricGroup": "Branches;PGO",
"MetricName": "BpTB"
},
{
"MetricExpr": "min( 1 , UOPS_ISSUED.ANY / ( (UOPS_RETIRED.RETIRE_SLOTS / INST_RETIRED.ANY) * 64 * ( ICACHE_64B.IFTAG_HIT + ICACHE_64B.IFTAG_MISS ) / 4.1 ) )",
"BriefDescription": "Rough Estimation of fraction of fetched lines bytes that were likely (includes speculatively fetches) consumed by program instructions",
"MetricGroup": "PGO",
"MetricName": "IFetch_Line_Utilization"
},
{
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded Icache; or Uop Cache)",
"MetricExpr": "IDQ.DSB_UOPS / ( IDQ.DSB_UOPS + LSD.UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS )",
"MetricGroup": "DSB; Frontend_Bandwidth",
"MetricExpr": "IDQ.DSB_UOPS / (( IDQ.DSB_UOPS + IDQ.MITE_UOPS + IDQ.MS_UOPS ))",
"BriefDescription": "Fraction of Uops delivered by the DSB (aka Decoded ICache; or Uop Cache)",
"MetricGroup": "DSB;Frontend_Bandwidth",
"MetricName": "DSB_Coverage"
},
{
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricExpr": "1 / (INST_RETIRED.ANY / cycles)",
"BriefDescription": "Cycles Per Instruction (threaded)",
"MetricGroup": "Pipeline;Summary",
"MetricName": "CPI"
},
{
"BriefDescription": "Per-thread actual clocks when the logical processor is active. This is called 'Clockticks' in VTune.",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD",
"BriefDescription": "Per-thread actual clocks when the logical processor is active.",
"MetricGroup": "Summary",
"MetricName": "CLKS"
},
{
"BriefDescription": "Total issue-pipeline slots",
"MetricExpr": "4*(( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricExpr": "4 * cycles",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1",
"MetricName": "SLOTS"
},
{
"BriefDescription": "Total number of retired Instructions",
"MetricExpr": "4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Total issue-pipeline slots (per core)",
"MetricGroup": "TopDownL1_SMT",
"MetricName": "SLOTS_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_INST_RETIRED.ALL_LOADS",
"BriefDescription": "Instructions per Load (lower number means loads are more frequent)",
"MetricGroup": "Instruction_Type;L1_Bound",
"MetricName": "IpL"
},
{
"MetricExpr": "INST_RETIRED.ANY / MEM_INST_RETIRED.ALL_STORES",
"BriefDescription": "Instructions per Store",
"MetricGroup": "Instruction_Type;Store_Bound",
"MetricName": "IpS"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.ALL_BRANCHES",
"BriefDescription": "Instructions per Branch",
"MetricGroup": "Branches;Instruction_Type;Port_5;Port_6",
"MetricName": "IpB"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_INST_RETIRED.NEAR_CALL",
"BriefDescription": "Instruction per (near) call",
"MetricGroup": "Branches",
"MetricName": "IpCall"
},
{
"MetricExpr": "INST_RETIRED.ANY",
"BriefDescription": "Total number of retired Instructions",
"MetricGroup": "Summary",
"MetricName": "Instructions"
},
{
"MetricExpr": "INST_RETIRED.ANY / cycles",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricExpr": "INST_RETIRED.ANY / (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC"
},
{
"MetricExpr": "INST_RETIRED.ANY / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Instructions Per Cycle (per physical core)",
"MetricGroup": "SMT",
"MetricName": "CoreIPC_SMT"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * ( FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE ) + 16 * FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE )) / cycles",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS",
"MetricName": "FLOPc"
},
{
"MetricExpr": "(( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * ( FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE ) + 16 * FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE )) / (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))",
"BriefDescription": "Floating Point Operations Per Cycle",
"MetricGroup": "FLOPS_SMT",
"MetricName": "FLOPc_SMT"
},
{
"MetricExpr": "UOPS_EXECUTED.THREAD / (( UOPS_EXECUTED.CORE_CYCLES_GE_1 / 2 ) if #SMT_on else UOPS_EXECUTED.CORE_CYCLES_GE_1)",
"BriefDescription": "Instruction-Level-Parallelism (average number of uops executed when there is at least 1 uop executed)",
"MetricExpr": "UOPS_EXECUTED.THREAD / (( UOPS_EXECUTED.CORE_CYCLES_GE_1 / 2) if #SMT_on else UOPS_EXECUTED.CORE_CYCLES_GE_1)",
"MetricGroup": "Pipeline;Ports_Utilization",
"MetricName": "ILP"
},
{
"BriefDescription": "Average Branch Address Clear Cost (fraction of cycles)",
"MetricExpr": "2* (( RS_EVENTS.EMPTY_CYCLES - ICACHE_16B.IFDATA_STALL - ICACHE_64B.IFTAG_STALL ) / RS_EVENTS.EMPTY_END)",
"MetricGroup": "Unknown_Branches",
"MetricName": "BAClear_Cost"
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * INT_MISC.RECOVERY_CYCLES ) / (4 * cycles))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) * (( INT_MISC.CLEAR_RESTEER_CYCLES + 9 * BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * cycles)) ) * (4 * cycles) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "Branch_Misprediction_Cost"
},
{
"MetricExpr": "( ((BR_MISP_RETIRED.ALL_BRANCHES / ( BR_MISP_RETIRED.ALL_BRANCHES + MACHINE_CLEARS.COUNT )) * (( UOPS_ISSUED.ANY - UOPS_RETIRED.RETIRE_SLOTS + 4 * (( INT_MISC.RECOVERY_CYCLES_ANY / 2 )) ) / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))))) + (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) * (( INT_MISC.CLEAR_RESTEER_CYCLES + 9 * BACLEARS.ANY ) / cycles) / (4 * IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE / (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )))) ) * (4 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) ))) / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Branch Misprediction Cost: Fraction of TopDown slots wasted per branch misprediction (jeclear and baclear)",
"MetricGroup": "Branch_Mispredicts_SMT",
"MetricName": "Branch_Misprediction_Cost_SMT"
},
{
"MetricExpr": "INST_RETIRED.ANY / BR_MISP_RETIRED.ALL_BRANCHES",
"BriefDescription": "Number of Instructions per non-speculative Branch Misprediction (JEClear)",
"MetricGroup": "Branch_Mispredicts",
"MetricName": "IpMispredict"
},
{
"MetricExpr": "( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )",
"BriefDescription": "Core actual clocks when any thread is active on the physical core",
"MetricExpr": "( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else CPU_CLK_UNHALTED.THREAD",
"MetricGroup": "SMT",
"MetricName": "CORE_CLKS"
},
{
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads",
"MetricExpr": "L1D_PEND_MISS.PENDING / ( MEM_LOAD_RETIRED.L1_MISS + MEM_LOAD_RETIRED.FB_HIT )",
"BriefDescription": "Actual Average Latency for L1 data-cache miss demand loads (in core cycles)",
"MetricGroup": "Memory_Bound;Memory_Lat",
"MetricName": "Load_Miss_Real_Latency"
},
{
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least 1 such miss)",
"MetricExpr": "L1D_PEND_MISS.PENDING / (( L1D_PEND_MISS.PENDING_CYCLES_ANY / 2) if #SMT_on else L1D_PEND_MISS.PENDING_CYCLES)",
"MetricExpr": "L1D_PEND_MISS.PENDING / L1D_PEND_MISS.PENDING_CYCLES",
"BriefDescription": "Memory-Level-Parallelism (average number of L1 miss demand load when there is at least one such miss. Per-thread)",
"MetricGroup": "Memory_Bound;Memory_BW",
"MetricName": "MLP"
},
{
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * cycles )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * (( CPU_CLK_UNHALTED.THREAD_ANY / 2 ) if #SMT_on else cycles) )",
"MetricGroup": "TLB",
"MetricName": "Page_Walks_Utilization"
},
{
"BriefDescription": "Average CPU Utilization",
"MetricExpr": "( ITLB_MISSES.WALK_PENDING + DTLB_LOAD_MISSES.WALK_PENDING + DTLB_STORE_MISSES.WALK_PENDING + EPT.WALK_PENDING ) / ( 2 * (( ( CPU_CLK_UNHALTED.THREAD / 2 ) * ( 1 + CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE / CPU_CLK_UNHALTED.REF_XCLK ) )) )",
"BriefDescription": "Utilization of the core's Page Walker(s) serving STLB misses triggered by instruction/Load/Store accesses",
"MetricGroup": "TLB_SMT",
"MetricName": "Page_Walks_Utilization_SMT"
},
{
"MetricExpr": "64 * L1D.REPLACEMENT / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L1 data cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L1D_Cache_Fill_BW"
},
{
"MetricExpr": "64 * L2_LINES_IN.ALL / 1000000000 / duration_time",
"BriefDescription": "Average data fill bandwidth to the L2 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L2_Cache_Fill_BW"
},
{
"MetricExpr": "64 * LONGEST_LAT_CACHE.MISS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Fill_BW"
},
{
"MetricExpr": "64 * OFFCORE_REQUESTS.ALL_REQUESTS / 1000000000 / duration_time",
"BriefDescription": "Average per-core data fill bandwidth to the L3 cache [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "L3_Cache_Access_BW"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L1_MISS / INST_RETIRED.ANY",
"BriefDescription": "L1 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L1MPKI"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L2_MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI"
},
{
"MetricExpr": "1000 * L2_RQSTS.MISS / INST_RETIRED.ANY",
"BriefDescription": "L2 cache misses per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2MPKI_All"
},
{
"MetricExpr": "1000 * ( L2_RQSTS.REFERENCES - L2_RQSTS.MISS ) / INST_RETIRED.ANY",
"BriefDescription": "L2 cache hits per kilo instruction for all request types (including speculative)",
"MetricGroup": "Cache_Misses;",
"MetricName": "L2HPKI_All"
},
{
"MetricExpr": "1000 * MEM_LOAD_RETIRED.L3_MISS / INST_RETIRED.ANY",
"BriefDescription": "L3 cache true misses per kilo instruction for retired demand loads",
"MetricGroup": "Cache_Misses;",
"MetricName": "L3MPKI"
},
{
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@",
"BriefDescription": "Average CPU Utilization",
"MetricGroup": "Summary",
"MetricName": "CPU_Utilization"
},
{
"MetricExpr": "( (( 1 * ( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2 * FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4 * ( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8 * ( FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE ) + 16 * FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE )) / 1000000000 ) / duration_time",
"BriefDescription": "Giga Floating Point Operations Per Second",
"MetricExpr": "(( 1*( FP_ARITH_INST_RETIRED.SCALAR_SINGLE + FP_ARITH_INST_RETIRED.SCALAR_DOUBLE ) + 2* FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE + 4*( FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE + FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE ) + 8* FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE )) / 1000000000 / duration_time",
"MetricGroup": "FLOPS;Summary",
"MetricName": "GFLOPs"
},
{
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricExpr": "CPU_CLK_UNHALTED.THREAD / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Average Frequency Utilization relative nominal frequency",
"MetricGroup": "Power",
"MetricName": "Turbo_Utilization"
},
{
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricExpr": "1 - CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE / ( CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY / 2 ) if #SMT_on else 0",
"BriefDescription": "Fraction of cycles where both hardware threads were active",
"MetricGroup": "SMT;Summary",
"MetricName": "SMT_2T_Utilization"
},
{
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricExpr": "CPU_CLK_UNHALTED.REF_TSC:u / CPU_CLK_UNHALTED.REF_TSC",
"BriefDescription": "Fraction of cycles spent in Kernel mode",
"MetricGroup": "Summary",
"MetricName": "Kernel_Utilization"
},
{
"BriefDescription": "C3 residency percent per core",
"MetricExpr": "( 64 * ( uncore_imc@cas_count_read@ + uncore_imc@cas_count_write@ ) / 1000000000 ) / duration_time",
"BriefDescription": "Average external Memory Bandwidth Use for reads and writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_BW_Use"
},
{
"MetricExpr": "1000000000 * ( cha@event\\=0x36\\\\\\,umask\\=0x21@ / cha@event\\=0x35\\\\\\,umask\\=0x21@ ) / ( cha_0@event\\=0x0@ / duration_time )",
"BriefDescription": "Average latency of data read request to external memory (in nanoseconds). Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_Lat",
"MetricName": "DRAM_Read_Latency"
},
{
"MetricExpr": "cha@event\\=0x36\\\\\\,umask\\=0x21@ / cha@event\\=0x36\\\\\\,umask\\=0x21\\\\\\,thresh\\=1@",
"BriefDescription": "Average number of parallel data read requests to external memory. Accounts for demand loads and L1/L2 prefetches",
"MetricGroup": "Memory_BW",
"MetricName": "DRAM_Parallel_Reads"
},
{
"MetricExpr": "( 1000000000 * ( imc@event\\=0xe0\\\\\\,umask\\=0x1@ / imc@event\\=0xe3@ ) / imc_0@event\\=0x0@ ) if 1 if 0 == 1 else 0 else 0",
"BriefDescription": "Average latency of data read request to external 3D X-Point memory [in nanoseconds]. Accounts for demand loads and L1/L2 data-read prefetches",
"MetricGroup": "Memory_Lat",
"MetricName": "MEM_PMM_Read_Latency"
},
{
"MetricExpr": "( ( 64 * imc@event\\=0xe3@ / 1000000000 ) / duration_time ) if 1 if 0 == 1 else 0 else 0",
"BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "PMM_Read_BW"
},
{
"MetricExpr": "( ( 64 * imc@event\\=0xe7@ / 1000000000 ) / duration_time ) if 1 if 0 == 1 else 0 else 0",
"BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes [GB / sec]",
"MetricGroup": "Memory_BW",
"MetricName": "PMM_Write_BW"
},
{
"MetricExpr": "cha_0@event\\=0x0@",
"BriefDescription": "Socket actual clocks when any core is active on that socket",
"MetricGroup": "",
"MetricName": "Socket_CLKS"
},
{
"MetricExpr": "(cstate_core@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per core",
"MetricName": "C3_Core_Residency"
},
{
"BriefDescription": "C6 residency percent per core",
"MetricExpr": "(cstate_core@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per core",
"MetricName": "C6_Core_Residency"
},
{
"BriefDescription": "C7 residency percent per core",
"MetricExpr": "(cstate_core@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per core",
"MetricName": "C7_Core_Residency"
},
{
"BriefDescription": "C2 residency percent per package",
"MetricExpr": "(cstate_pkg@c2\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C2 residency percent per package",
"MetricName": "C2_Pkg_Residency"
},
{
"BriefDescription": "C3 residency percent per package",
"MetricExpr": "(cstate_pkg@c3\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C3 residency percent per package",
"MetricName": "C3_Pkg_Residency"
},
{
"BriefDescription": "C6 residency percent per package",
"MetricExpr": "(cstate_pkg@c6\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C6 residency percent per package",
"MetricName": "C6_Pkg_Residency"
},
{
"BriefDescription": "C7 residency percent per package",
"MetricExpr": "(cstate_pkg@c7\\-residency@ / msr@tsc@) * 100",
"MetricGroup": "Power",
"BriefDescription": "C7 residency percent per package",
"MetricName": "C7_Pkg_Residency"
}
]

1
tools/perf/trace/beauty/renameat.c
View file

@ -2,7 +2,6 @@
// Copyright (C) 2018, Red Hat Inc, Arnaldo Carvalho de Melo <acme@redhat.com>
#include "trace/beauty/beauty.h"
#include <uapi/linux/fs.h>
static size_t renameat2__scnprintf_flags(unsigned long flags, char *bf, size_t size, bool show_prefix)
{

Some files were not shown because too many files have changed in this diff Show more