mirror of
https://github.com/jart/cosmopolitan.git
synced 2025-05-22 21:32:31 +00:00
Perform build and magnum tuning
Building o//third_party/python now takes 5 seconds on my PC This change works towards modifying Python to use runtime dispatching when appropriate. For example, when loading the magnums in the socket module, it's a good idea to check if the magnum is zero, because that means the local system platform doesn't support it.
This commit is contained in:
parent
ee7e296339
commit
d26d7ae0e4
1028 changed files with 6576 additions and 172777 deletions
135
third_party/python/Objects/stringlib/find_max_char.inc
vendored
Normal file
135
third_party/python/Objects/stringlib/find_max_char.inc
vendored
Normal file
|
@ -0,0 +1,135 @@
|
|||
/* clang-format off */
|
||||
/* Finding the optimal width of unicode characters in a buffer */
|
||||
|
||||
#if !STRINGLIB_IS_UNICODE
|
||||
# error "find_max_char.h is specific to Unicode"
|
||||
#endif
|
||||
|
||||
/* Mask to quickly check whether a C 'long' contains a
|
||||
non-ASCII, UTF8-encoded char. */
|
||||
#if (SIZEOF_LONG == 8)
|
||||
# define UCS1_ASCII_CHAR_MASK 0x8080808080808080UL
|
||||
#elif (SIZEOF_LONG == 4)
|
||||
# define UCS1_ASCII_CHAR_MASK 0x80808080UL
|
||||
#else
|
||||
# error C 'long' size should be either 4 or 8!
|
||||
#endif
|
||||
|
||||
#if STRINGLIB_SIZEOF_CHAR == 1
|
||||
|
||||
Py_LOCAL_INLINE(Py_UCS4)
|
||||
STRINGLIB(find_max_char)(const STRINGLIB_CHAR *begin, const STRINGLIB_CHAR *end)
|
||||
{
|
||||
const unsigned char *p = (const unsigned char *) begin;
|
||||
const unsigned char *aligned_end =
|
||||
(const unsigned char *) _Py_ALIGN_DOWN(end, SIZEOF_LONG);
|
||||
|
||||
while (p < end) {
|
||||
if (_Py_IS_ALIGNED(p, SIZEOF_LONG)) {
|
||||
/* Help register allocation */
|
||||
const unsigned char *_p = p;
|
||||
while (_p < aligned_end) {
|
||||
unsigned long value = *(unsigned long *) _p;
|
||||
if (value & UCS1_ASCII_CHAR_MASK)
|
||||
return 255;
|
||||
_p += SIZEOF_LONG;
|
||||
}
|
||||
p = _p;
|
||||
if (p == end)
|
||||
break;
|
||||
}
|
||||
if (*p++ & 0x80)
|
||||
return 255;
|
||||
}
|
||||
return 127;
|
||||
}
|
||||
|
||||
#undef ASCII_CHAR_MASK
|
||||
|
||||
#else /* STRINGLIB_SIZEOF_CHAR == 1 */
|
||||
|
||||
#define MASK_ASCII 0xFFFFFF80
|
||||
#define MASK_UCS1 0xFFFFFF00
|
||||
#define MASK_UCS2 0xFFFF0000
|
||||
|
||||
#define MAX_CHAR_ASCII 0x7f
|
||||
#define MAX_CHAR_UCS1 0xff
|
||||
#define MAX_CHAR_UCS2 0xffff
|
||||
#define MAX_CHAR_UCS4 0x10ffff
|
||||
|
||||
Py_LOCAL_INLINE(Py_UCS4)
|
||||
STRINGLIB(find_max_char)(const STRINGLIB_CHAR *begin, const STRINGLIB_CHAR *end)
|
||||
{
|
||||
#if STRINGLIB_SIZEOF_CHAR == 2
|
||||
const Py_UCS4 mask_limit = MASK_UCS1;
|
||||
const Py_UCS4 max_char_limit = MAX_CHAR_UCS2;
|
||||
#elif STRINGLIB_SIZEOF_CHAR == 4
|
||||
const Py_UCS4 mask_limit = MASK_UCS2;
|
||||
const Py_UCS4 max_char_limit = MAX_CHAR_UCS4;
|
||||
#else
|
||||
#error Invalid STRINGLIB_SIZEOF_CHAR (must be 1, 2 or 4)
|
||||
#endif
|
||||
Py_UCS4 mask;
|
||||
Py_ssize_t n = end - begin;
|
||||
const STRINGLIB_CHAR *p = begin;
|
||||
const STRINGLIB_CHAR *unrolled_end = begin + _Py_SIZE_ROUND_DOWN(n, 4);
|
||||
Py_UCS4 max_char;
|
||||
|
||||
max_char = MAX_CHAR_ASCII;
|
||||
mask = MASK_ASCII;
|
||||
while (p < unrolled_end) {
|
||||
STRINGLIB_CHAR bits = p[0] | p[1] | p[2] | p[3];
|
||||
if (bits & mask) {
|
||||
if (mask == mask_limit) {
|
||||
/* Limit reached */
|
||||
return max_char_limit;
|
||||
}
|
||||
if (mask == MASK_ASCII) {
|
||||
max_char = MAX_CHAR_UCS1;
|
||||
mask = MASK_UCS1;
|
||||
}
|
||||
else {
|
||||
/* mask can't be MASK_UCS2 because of mask_limit above */
|
||||
assert(mask == MASK_UCS1);
|
||||
max_char = MAX_CHAR_UCS2;
|
||||
mask = MASK_UCS2;
|
||||
}
|
||||
/* We check the new mask on the same chars in the next iteration */
|
||||
continue;
|
||||
}
|
||||
p += 4;
|
||||
}
|
||||
while (p < end) {
|
||||
if (p[0] & mask) {
|
||||
if (mask == mask_limit) {
|
||||
/* Limit reached */
|
||||
return max_char_limit;
|
||||
}
|
||||
if (mask == MASK_ASCII) {
|
||||
max_char = MAX_CHAR_UCS1;
|
||||
mask = MASK_UCS1;
|
||||
}
|
||||
else {
|
||||
/* mask can't be MASK_UCS2 because of mask_limit above */
|
||||
assert(mask == MASK_UCS1);
|
||||
max_char = MAX_CHAR_UCS2;
|
||||
mask = MASK_UCS2;
|
||||
}
|
||||
/* We check the new mask on the same chars in the next iteration */
|
||||
continue;
|
||||
}
|
||||
p++;
|
||||
}
|
||||
return max_char;
|
||||
}
|
||||
|
||||
#undef MASK_ASCII
|
||||
#undef MASK_UCS1
|
||||
#undef MASK_UCS2
|
||||
#undef MAX_CHAR_ASCII
|
||||
#undef MAX_CHAR_UCS1
|
||||
#undef MAX_CHAR_UCS2
|
||||
#undef MAX_CHAR_UCS4
|
||||
|
||||
#endif /* STRINGLIB_SIZEOF_CHAR == 1 */
|
||||
|
Loading…
Add table
Add a link
Reference in a new issue