mirrors/cpython
1
0
Fork 0
mirror of https://github.com/python/cpython.git synced 2025-09-27 10:50:04 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 13

GH-115685: Optimize TO_BOOL and variants based on truthiness of input. (GH-116311)

Browse source
This commit is contained in:
Mark Shannon 2024-03-05 11:23:46 +00:00 • committed by GitHub
parent a29998a06b
commit cbf3d38cbe
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 150 additions and 55 deletions
Download patch file Download diff file Expand all files Collapse all files

1
Include/internal/pycore_optimizer.h
View file

@ -96,6 +96,7 @@ extern bool _Py_uop_sym_set_non_null(_Py_UopsSymbol *sym);
extern bool _Py_uop_sym_set_type(_Py_UopsSymbol *sym, PyTypeObject *typ); extern bool _Py_uop_sym_set_type(_Py_UopsSymbol *sym, PyTypeObject *typ);
extern bool _Py_uop_sym_set_const(_Py_UopsSymbol *sym, PyObject *const_val); extern bool _Py_uop_sym_set_const(_Py_UopsSymbol *sym, PyObject *const_val);
extern bool _Py_uop_sym_is_bottom(_Py_UopsSymbol *sym); extern bool _Py_uop_sym_is_bottom(_Py_UopsSymbol *sym);
extern int _Py_uop_sym_truthiness(_Py_UopsSymbol *sym);
extern int _Py_uop_abstractcontext_init(_Py_UOpsContext *ctx); extern int _Py_uop_abstractcontext_init(_Py_UOpsContext *ctx);

22
Python/optimizer_analysis.c
View file

@ -298,9 +298,31 @@ remove_globals(_PyInterpreterFrame *frame, _PyUOpInstruction *buffer,
#define sym_set_type _Py_uop_sym_set_type #define sym_set_type _Py_uop_sym_set_type
#define sym_set_const _Py_uop_sym_set_const #define sym_set_const _Py_uop_sym_set_const
#define sym_is_bottom _Py_uop_sym_is_bottom #define sym_is_bottom _Py_uop_sym_is_bottom
#define sym_truthiness _Py_uop_sym_truthiness
#define frame_new _Py_uop_frame_new #define frame_new _Py_uop_frame_new
#define frame_pop _Py_uop_frame_pop #define frame_pop _Py_uop_frame_pop
static int
optimize_to_bool(
_PyUOpInstruction *this_instr,
_Py_UOpsContext *ctx,
_Py_UopsSymbol *value,
_Py_UopsSymbol **result_ptr)
{
if (sym_matches_type(value, &PyBool_Type)) {
REPLACE_OP(this_instr, _NOP, 0, 0);
*result_ptr = value;
return 1;
}
int truthiness = sym_truthiness(value);
if (truthiness >= 0) {
PyObject *load = truthiness ? Py_True : Py_False;
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)load);
*result_ptr = sym_new_const(ctx, load);
return 1;
}
return 0;
}
/* 1 for success, 0 for not ready, cannot error at the moment. */ /* 1 for success, 0 for not ready, cannot error at the moment. */
static int static int

73
Python/optimizer_bytecodes.c
View file

@ -29,6 +29,14 @@ typedef struct _Py_UOpsAbstractFrame _Py_UOpsAbstractFrame;
#define frame_new _Py_uop_frame_new #define frame_new _Py_uop_frame_new
#define frame_pop _Py_uop_frame_pop #define frame_pop _Py_uop_frame_pop
extern int
optimize_to_bool(
_PyUOpInstruction *this_instr,
_Py_UOpsContext *ctx,
_Py_UopsSymbol *value,
_Py_UopsSymbol **result_ptr);
static int static int
dummy_func(void) { dummy_func(void) {
@ -271,63 +279,72 @@ dummy_func(void) {
} }
op(_TO_BOOL, (value -- res)) { op(_TO_BOOL, (value -- res)) {
(void)value; if (optimize_to_bool(this_instr, ctx, value, &res)) {
res = sym_new_type(ctx, &PyBool_Type); OUT_OF_SPACE_IF_NULL(res);
OUT_OF_SPACE_IF_NULL(res); }
else {
res = sym_new_type(ctx, &PyBool_Type);
OUT_OF_SPACE_IF_NULL(res);
}
} }
op(_TO_BOOL_BOOL, (value -- value)) { op(_TO_BOOL_BOOL, (value -- res)) {
if (sym_matches_type(value, &PyBool_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
REPLACE_OP(this_instr, _NOP, 0, 0); OUT_OF_SPACE_IF_NULL(res);
} }
else { else {
if(!sym_set_type(value, &PyBool_Type)) { if(!sym_set_type(value, &PyBool_Type)) {
goto hit_bottom; goto hit_bottom;
} }
res = value;
} }
} }
op(_TO_BOOL_INT, (value -- res)) { op(_TO_BOOL_INT, (value -- res)) {
if (sym_is_const(value) && sym_matches_type(value, &PyLong_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
PyObject *load = _PyLong_IsZero((PyLongObject *)sym_get_const(value)) OUT_OF_SPACE_IF_NULL(res);
? Py_False : Py_True;
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)load);
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, load));
} }
else { else {
if(!sym_set_type(value, &PyLong_Type)) {
goto hit_bottom;
}
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type)); OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} }
if(!sym_set_type(value, &PyLong_Type)) {
goto hit_bottom;
}
} }
op(_TO_BOOL_LIST, (value -- res)) { op(_TO_BOOL_LIST, (value -- res)) {
if(!sym_set_type(value, &PyList_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
goto hit_bottom; OUT_OF_SPACE_IF_NULL(res);
}
else {
if(!sym_set_type(value, &PyList_Type)) {
goto hit_bottom;
}
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} }
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} }
op(_TO_BOOL_NONE, (value -- res)) { op(_TO_BOOL_NONE, (value -- res)) {
if (sym_get_const(value) == Py_None) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)Py_False); OUT_OF_SPACE_IF_NULL(res);
}
else {
if (!sym_set_const(value, Py_None)) {
goto hit_bottom;
}
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, Py_False));
} }
sym_set_const(value, Py_None);
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, Py_False));
} }
op(_TO_BOOL_STR, (value -- res)) { op(_TO_BOOL_STR, (value -- res)) {
if (sym_is_const(value) && sym_matches_type(value, &PyUnicode_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
PyObject *load = sym_get_const(value) == &_Py_STR(empty) ? Py_False : Py_True; OUT_OF_SPACE_IF_NULL(res);
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)load);
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, load));
} }
else { else {
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type)); OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} if(!sym_set_type(value, &PyUnicode_Type)) {
if(!sym_set_type(value, &PyUnicode_Type)) { goto hit_bottom;
goto hit_bottom; }
} }
} }

65
Python/optimizer_cases.c.h generated
View file

@ -107,24 +107,31 @@
_Py_UopsSymbol *value; _Py_UopsSymbol *value;
_Py_UopsSymbol *res; _Py_UopsSymbol *res;
value = stack_pointer[-1]; value = stack_pointer[-1];
(void)value; if (optimize_to_bool(this_instr, ctx, value, &res)) {
res = sym_new_type(ctx, &PyBool_Type); OUT_OF_SPACE_IF_NULL(res);
OUT_OF_SPACE_IF_NULL(res); }
else {
res = sym_new_type(ctx, &PyBool_Type);
OUT_OF_SPACE_IF_NULL(res);
}
stack_pointer[-1] = res; stack_pointer[-1] = res;
break; break;
} }
case _TO_BOOL_BOOL: { case _TO_BOOL_BOOL: {
_Py_UopsSymbol *value; _Py_UopsSymbol *value;
_Py_UopsSymbol *res;
value = stack_pointer[-1]; value = stack_pointer[-1];
if (sym_matches_type(value, &PyBool_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
REPLACE_OP(this_instr, _NOP, 0, 0); OUT_OF_SPACE_IF_NULL(res);
} }
else { else {
if(!sym_set_type(value, &PyBool_Type)) { if(!sym_set_type(value, &PyBool_Type)) {
goto hit_bottom; goto hit_bottom;
} }
res = value;
} }
stack_pointer[-1] = res;
break; break;
} }
@ -132,18 +139,15 @@
_Py_UopsSymbol *value; _Py_UopsSymbol *value;
_Py_UopsSymbol *res; _Py_UopsSymbol *res;
value = stack_pointer[-1]; value = stack_pointer[-1];
if (sym_is_const(value) && sym_matches_type(value, &PyLong_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
PyObject *load = _PyLong_IsZero((PyLongObject *)sym_get_const(value)) OUT_OF_SPACE_IF_NULL(res);
? Py_False : Py_True;
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)load);
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, load));
} }
else { else {
if(!sym_set_type(value, &PyLong_Type)) {
goto hit_bottom;
}
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type)); OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} }
if(!sym_set_type(value, &PyLong_Type)) {
goto hit_bottom;
}
stack_pointer[-1] = res; stack_pointer[-1] = res;
break; break;
} }
@ -152,10 +156,15 @@
_Py_UopsSymbol *value; _Py_UopsSymbol *value;
_Py_UopsSymbol *res; _Py_UopsSymbol *res;
value = stack_pointer[-1]; value = stack_pointer[-1];
if(!sym_set_type(value, &PyList_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
goto hit_bottom; OUT_OF_SPACE_IF_NULL(res);
}
else {
if(!sym_set_type(value, &PyList_Type)) {
goto hit_bottom;
}
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} }
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
stack_pointer[-1] = res; stack_pointer[-1] = res;
break; break;
} }
@ -164,11 +173,15 @@
_Py_UopsSymbol *value; _Py_UopsSymbol *value;
_Py_UopsSymbol *res; _Py_UopsSymbol *res;
value = stack_pointer[-1]; value = stack_pointer[-1];
if (sym_get_const(value) == Py_None) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)Py_False); OUT_OF_SPACE_IF_NULL(res);
}
else {
if (!sym_set_const(value, Py_None)) {
goto hit_bottom;
}
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, Py_False));
} }
sym_set_const(value, Py_None);
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, Py_False));
stack_pointer[-1] = res; stack_pointer[-1] = res;
break; break;
} }
@ -177,16 +190,14 @@
_Py_UopsSymbol *value; _Py_UopsSymbol *value;
_Py_UopsSymbol *res; _Py_UopsSymbol *res;
value = stack_pointer[-1]; value = stack_pointer[-1];
if (sym_is_const(value) && sym_matches_type(value, &PyUnicode_Type)) { if (optimize_to_bool(this_instr, ctx, value, &res)) {
PyObject *load = sym_get_const(value) == &_Py_STR(empty) ? Py_False : Py_True; OUT_OF_SPACE_IF_NULL(res);
REPLACE_OP(this_instr, _POP_TOP_LOAD_CONST_INLINE_BORROW, 0, (uintptr_t)load);
OUT_OF_SPACE_IF_NULL(res = sym_new_const(ctx, load));
} }
else { else {
OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type)); OUT_OF_SPACE_IF_NULL(res = sym_new_type(ctx, &PyBool_Type));
} if(!sym_set_type(value, &PyUnicode_Type)) {
if(!sym_set_type(value, &PyUnicode_Type)) { goto hit_bottom;
goto hit_bottom; }
} }
stack_pointer[-1] = res; stack_pointer[-1] = res;
break; break;

44
Python/optimizer_symbols.c
View file

@ -4,6 +4,7 @@
#include "cpython/optimizer.h" #include "cpython/optimizer.h"
#include "pycore_code.h" #include "pycore_code.h"
#include "pycore_frame.h" #include "pycore_frame.h"
#include "pycore_long.h"
#include "pycore_optimizer.h" #include "pycore_optimizer.h"
#include <stdbool.h> #include <stdbool.h>
@ -240,6 +241,40 @@ _Py_uop_sym_matches_type(_Py_UopsSymbol *sym, PyTypeObject *typ)
return sym->typ == typ; return sym->typ == typ;
} }
int
_Py_uop_sym_truthiness(_Py_UopsSymbol *sym)
{
/* There are some non-constant values for
* which `bool(val)` always evaluates to
* True or False, such as tuples with known
* length, but unknown contents, or bound-methods.
* This function will need updating
* should we support those values.
*/
if (_Py_uop_sym_is_bottom(sym)) {
return -1;
}
if (!_Py_uop_sym_is_const(sym)) {
return -1;
}
PyObject *value = _Py_uop_sym_get_const(sym);
if (value == Py_None) {
return 0;
}
/* Only handle a few known safe types */
PyTypeObject *tp = Py_TYPE(value);
if (tp == &PyLong_Type) {
return !_PyLong_IsZero((PyLongObject *)value);
}
if (tp == &PyUnicode_Type) {
return value != &_Py_STR(empty);
}
if (tp == &PyBool_Type) {
return value == Py_True;
}
return -1;
}
// 0 on success, -1 on error. // 0 on success, -1 on error.
_Py_UOpsAbstractFrame * _Py_UOpsAbstractFrame *
_Py_uop_frame_new( _Py_uop_frame_new(
@ -413,6 +448,7 @@ _Py_uop_symbols_test(PyObject *Py_UNUSED(self), PyObject *Py_UNUSED(ignored))
goto fail; goto fail;
} }
_Py_uop_sym_set_const(sym, val_42); _Py_uop_sym_set_const(sym, val_42);
TEST_PREDICATE(_Py_uop_sym_truthiness(sym) == 1, "bool(42) is not True");
TEST_PREDICATE(!_Py_uop_sym_is_null(sym), "42 is NULL"); TEST_PREDICATE(!_Py_uop_sym_is_null(sym), "42 is NULL");
TEST_PREDICATE(_Py_uop_sym_is_not_null(sym), "42 isn't not NULL"); TEST_PREDICATE(_Py_uop_sym_is_not_null(sym), "42 isn't not NULL");
TEST_PREDICATE(_Py_uop_sym_matches_type(sym, &PyLong_Type), "42 isn't an int"); TEST_PREDICATE(_Py_uop_sym_matches_type(sym, &PyLong_Type), "42 isn't an int");
@ -436,6 +472,14 @@ _Py_uop_symbols_test(PyObject *Py_UNUSED(self), PyObject *Py_UNUSED(ignored))
_Py_uop_sym_set_const(sym, val_43); // Should make it bottom _Py_uop_sym_set_const(sym, val_43); // Should make it bottom
TEST_PREDICATE(_Py_uop_sym_is_bottom(sym), "(42 and 43) isn't bottom"); TEST_PREDICATE(_Py_uop_sym_is_bottom(sym), "(42 and 43) isn't bottom");
sym = _Py_uop_sym_new_const(ctx, Py_None);
TEST_PREDICATE(_Py_uop_sym_truthiness(sym) == 0, "bool(None) is not False");
sym = _Py_uop_sym_new_const(ctx, Py_False);
TEST_PREDICATE(_Py_uop_sym_truthiness(sym) == 0, "bool(False) is not False");
sym = _Py_uop_sym_new_const(ctx, PyLong_FromLong(0));
TEST_PREDICATE(_Py_uop_sym_truthiness(sym) == 0, "bool(0) is not False");
_Py_uop_abstractcontext_fini(ctx); _Py_uop_abstractcontext_fini(ctx);
Py_DECREF(val_42); Py_DECREF(val_42);
Py_DECREF(val_43); Py_DECREF(val_43);