vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

rename device_infos to infos

Browse source
This commit is contained in:
arthw 2024-08-01 12:48:18 +08:00
parent 6211ac0408
commit 254a750249
5 changed files with 41 additions and 44 deletions
Download patch file Download diff file Expand all files Collapse all files

16
ggml/src/ggml-sycl.cpp
View file

@ -2280,11 +2280,11 @@ static int64_t get_row_rounding(ggml_type type, const std::array<float, GGML_SYC
for (int i = 0; i < ggml_sycl_info().device_count; ++i) { for (int i = 0; i < ggml_sycl_info().device_count; ++i) {
int id = ggml_backend_sycl_get_device_id(i); int id = ggml_backend_sycl_get_device_id(i);
if (tensor_split[i] < (i + 1 < ggml_sycl_info().device_count ? tensor_split[i + 1] : 1.0f)) { if (tensor_split[i] < (i + 1 < ggml_sycl_info().device_count ? tensor_split[i + 1] : 1.0f)) {
if (min_compute_capability > ggml_sycl_info().device_infos[id].cc) { if (min_compute_capability > ggml_sycl_info().infos[id].cc) {
min_compute_capability = ggml_sycl_info().device_infos[id].cc; min_compute_capability = ggml_sycl_info().infos[id].cc;
} }
if (max_compute_capability < ggml_sycl_info().device_infos[id].cc) { if (max_compute_capability < ggml_sycl_info().infos[id].cc) {
max_compute_capability = ggml_sycl_info().device_infos[id].cc; max_compute_capability = ggml_sycl_info().infos[id].cc;
} }
} }
} }
@ -3416,12 +3416,12 @@ static void ggml_sycl_mul_mat(ggml_backend_sycl_context & ctx, const ggml_tensor
continue; continue;
} }
if (min_compute_capability > ggml_sycl_info().device_infos[id].cc) { if (min_compute_capability > ggml_sycl_info().infos[id].cc) {
min_compute_capability = ggml_sycl_info().device_infos[id].cc; min_compute_capability = ggml_sycl_info().infos[id].cc;
} }
} }
} else { } else {
min_compute_capability = ggml_sycl_info().device_infos[ctx.device].cc; min_compute_capability = ggml_sycl_info().infos[ctx.device].cc;
} }
// check data types and tensor shapes for custom matrix multiplication kernels: // check data types and tensor shapes for custom matrix multiplication kernels:
@ -4342,7 +4342,7 @@ ggml_backend_buffer_type_t ggml_backend_sycl_buffer_type(int device_id) {
if (!ggml_backend_sycl_buffer_type_initialized) { if (!ggml_backend_sycl_buffer_type_initialized) {
for (auto & id: ggml_sycl_info().ids) { for (auto & id: ggml_sycl_info().ids) {
auto & device = dpct::dev_mgr::instance().get_device(id); auto & device = dpct::dev_mgr::instance().get_device(id);
queue_ptr stream = ggml_sycl_info().device_infos[id].qptrs[0]; queue_ptr stream = ggml_sycl_info().infos[id].qptrs[0];
ggml_backend_sycl_buffer_types[id] = { ggml_backend_sycl_buffer_types[id] = {
/* .iface = */ ggml_backend_sycl_buffer_type_interface, /* .iface = */ ggml_backend_sycl_buffer_type_interface,
/* .context = */ new ggml_backend_sycl_buffer_type_context{id, GGML_SYCL_NAME + std::to_string(id), stream}, /* .context = */ new ggml_backend_sycl_buffer_type_context{id, GGML_SYCL_NAME + std::to_string(id), stream},

2
ggml/src/ggml-sycl/common.hpp
View file

@ -241,7 +241,7 @@ struct ggml_backend_sycl_context {
device(id), device(id),
name(GGML_SYCL_NAME + std::to_string(device)) { name(GGML_SYCL_NAME + std::to_string(device)) {
for (int i=0;i<GGML_SYCL_MAX_STREAMS; i++){ for (int i=0;i<GGML_SYCL_MAX_STREAMS; i++){
qptrs[id][i] = sycl_device_info.device_infos[id].qptrs[i]; qptrs[id][i] = sycl_device_info.infos[id].qptrs[i];
} }
} }

20
ggml/src/ggml-sycl/mmq.cpp
View file

@ -1779,7 +1779,7 @@ static void ggml_mul_mat_q4_0_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -1894,7 +1894,7 @@ static void ggml_mul_mat_q4_1_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2009,7 +2009,7 @@ static void ggml_mul_mat_q5_0_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2124,7 +2124,7 @@ static void ggml_mul_mat_q5_1_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2239,7 +2239,7 @@ static void ggml_mul_mat_q8_0_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2354,7 +2354,7 @@ static void ggml_mul_mat_q2_K_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2477,7 +2477,7 @@ static void ggml_mul_mat_q3_K_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2605,7 +2605,7 @@ static void ggml_mul_mat_q4_K_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2726,7 +2726,7 @@ static void ggml_mul_mat_q5_K_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {
@ -2847,7 +2847,7 @@ static void ggml_mul_mat_q6_K_q8_1_sycl(const void *vx, const void *vy,
int id; int id;
SYCL_CHECK( SYCL_CHECK(
CHECK_TRY_ERROR(id = get_current_device_id())); CHECK_TRY_ERROR(id = get_current_device_id()));
const int compute_capability = ggml_sycl_info().device_infos[id].cc; const int compute_capability = ggml_sycl_info().infos[id].cc;
int mmq_x, mmq_y, nwarps; int mmq_x, mmq_y, nwarps;
if (compute_capability >= VER_GEN13) { if (compute_capability >= VER_GEN13) {

43
ggml/src/ggml-sycl/sycl_device.cpp
View file

@ -22,18 +22,15 @@ void ggml_sycl_device_info::init(
m_device_filter = device_filter; m_device_filter = device_filter;
} }
void ggml_sycl_device_info::clear_device_infos() { void ggml_sycl_device_info::clear_infos() {
ids.clear(); ids.clear();
devices.clear(); devices.clear();
for (int id=0;id<GGML_SYCL_MAX_DEVICES;id++) { for (int id=0;id<GGML_SYCL_MAX_DEVICES;id++) {
device_infos[id].id = -1; infos[id].id = -1;
device_infos[id].max_work_group_sizes = 0; infos[id].max_work_group_sizes = 0;
device_infos[id].max_compute_units = 0; infos[id].max_compute_units = 0;
device_infos[id].hw_family = -1; infos[id].hw_family = -1;
// for (int i=0; i<GGML_SYCL_MAX_STREAMS;i++) {
// free(device_infos[id].qptrs[i]);
// }
} }
device_count = 0; device_count = 0;
@ -43,7 +40,7 @@ void ggml_sycl_device_info::clear_device_infos() {
void ggml_sycl_device_info::init_single_mode(int main_gpu_id) { void ggml_sycl_device_info::init_single_mode(int main_gpu_id) {
GGML_ASSERT(main_gpu_id<dpct::dev_mgr::instance().device_count()); GGML_ASSERT(main_gpu_id<dpct::dev_mgr::instance().device_count());
clear_device_infos(); clear_infos();
add_device_info(main_gpu_id); add_device_info(main_gpu_id);
init_devices_dynamic_info(); init_devices_dynamic_info();
device_mode = SYCL_SINGLE_GPU_MODE; device_mode = SYCL_SINGLE_GPU_MODE;
@ -217,23 +214,23 @@ void ggml_sycl_device_info::add_device_info(int id) {
ids.push_back(id); ids.push_back(id);
devices.push_back(device); devices.push_back(device);
device_infos[id].id = id; infos[id].id = id;
device_infos[id].device = device; infos[id].device = device;
device_infos[id].max_work_group_sizes = prop.get_max_work_group_size(); infos[id].max_work_group_sizes = prop.get_max_work_group_size();
device_infos[id].max_compute_units = prop.get_max_compute_units(); infos[id].max_compute_units = prop.get_max_compute_units();
device_infos[id].hw_family = get_device_family(&device); infos[id].hw_family = get_device_family(&device);
} }
void ggml_sycl_device_info::create_queues(int id) { void ggml_sycl_device_info::create_queues(int id) {
for (int i=0; i<GGML_SYCL_MAX_STREAMS;i++) { for (int i=0; i<GGML_SYCL_MAX_STREAMS;i++) {
device_infos[id].qptrs[i] = create_queue_for_device_id(id); infos[id].qptrs[i] = create_queue_for_device_id(id);
} }
} }
void ggml_sycl_device_info::create_queues_for_devices() { void ggml_sycl_device_info::create_queues_for_devices() {
for (auto &id: ids) { for (auto &id: ids) {
for (int i=0; i<GGML_SYCL_MAX_STREAMS;i++) { for (int i=0; i<GGML_SYCL_MAX_STREAMS;i++) {
device_infos[id].qptrs[i] = create_queue_for_device_id(id); infos[id].qptrs[i] = create_queue_for_device_id(id);
} }
} }
} }
@ -259,28 +256,28 @@ void ggml_sycl_device_info::print_gpu_device_list() {
"units:%d, to use any SYCL devices, set/export " "units:%d, to use any SYCL devices, set/export "
"GGML_SYCL_VISIBLE_DEVICES or ONEAPI_DEVICE_SELECTOR\n"; "GGML_SYCL_VISIBLE_DEVICES or ONEAPI_DEVICE_SELECTOR\n";
fprintf(stderr, hint, get_device_count(), devices_list(), fprintf(stderr, hint, get_device_count(), devices_list(),
device_infos[0].max_compute_units); infos[0].max_compute_units);
} }
} }
int ggml_sycl_device_info::work_group_size(int id) { int ggml_sycl_device_info::work_group_size(int id) {
GGML_ASSERT(is_allowed_device(id)); GGML_ASSERT(is_allowed_device(id));
return device_infos[id].max_work_group_sizes; return infos[id].max_work_group_sizes;
} }
void ggml_sycl_device_info::update_mem() { void ggml_sycl_device_info::update_mem() {
for (int i = 0; i < GGML_SYCL_MAX_DEVICES; ++i) { for (int i = 0; i < GGML_SYCL_MAX_DEVICES; ++i) {
device_infos[i].vmm = 0; infos[i].vmm = 0;
default_tensor_split[i] = 0; default_tensor_split[i] = 0;
device_infos[i].cc =0; infos[i].cc =0;
} }
int64_t total_vram = 0; int64_t total_vram = 0;
for (int i = 0; i < device_count; ++i) { for (int i = 0; i < device_count; ++i) {
int id = get_device_id(i); int id = get_device_id(i);
device_infos[id].vmm = 0; infos[id].vmm = 0;
dpct::device_info prop; dpct::device_info prop;
dpct::get_device_info( dpct::get_device_info(
prop, dpct::dev_mgr::instance().get_device(id)); prop, dpct::dev_mgr::instance().get_device(id));
@ -289,7 +286,7 @@ void ggml_sycl_device_info::update_mem() {
default_tensor_split[i] = total_vram; default_tensor_split[i] = total_vram;
total_vram += prop.get_global_mem_size(); total_vram += prop.get_global_mem_size();
device_infos[id].cc = infos[id].cc =
100 * prop.get_major_version() + 10 * prop.get_minor_version(); 100 * prop.get_major_version() + 10 * prop.get_minor_version();
} }
@ -335,7 +332,7 @@ int ggml_sycl_device_info::get_device_id(int device_index) {
} }
int ggml_sycl_device_info::hw_family(int id) { int ggml_sycl_device_info::hw_family(int id) {
return device_infos[id].hw_family; return infos[id].hw_family;
} }
static inline bool env_existed(const char *env_name) { static inline bool env_existed(const char *env_name) {

4
ggml/src/ggml-sycl/sycl_device.hpp
View file

@ -54,7 +54,7 @@ struct ggml_sycl_device_info {
sycl::context co_ctx; sycl::context co_ctx;
int m_device_filter; int m_device_filter;
sycl_device_info device_infos[GGML_SYCL_MAX_DEVICES]; sycl_device_info infos[GGML_SYCL_MAX_DEVICES];
std::array<float, GGML_SYCL_MAX_DEVICES> default_tensor_split = {}; std::array<float, GGML_SYCL_MAX_DEVICES> default_tensor_split = {};
ggml_sycl_device_info(int main_gpu_id);//single device mode ggml_sycl_device_info(int main_gpu_id);//single device mode
@ -62,7 +62,7 @@ struct ggml_sycl_device_info {
void init(ggml_sycl_backend_device_filter device_filter); void init(ggml_sycl_backend_device_filter device_filter);
void init_single_mode(int main_gpu_id); void init_single_mode(int main_gpu_id);
void clear_device_infos(); void clear_infos();
void print_gpu_device_list(); void print_gpu_device_list();
int work_group_size(int device_id); int work_group_size(int device_id);
bool is_allowed_device(int device_id); bool is_allowed_device(int device_id);