Skip to content

gh-134584: Eliminate redundant refcounting from _BINARY_OP_ADD_UNICODE#135817

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
corona10 wants to merge 5 commits into from
Closed

gh-134584: Eliminate redundant refcounting from _BINARY_OP_ADD_UNICODE #135817

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
d5afe15
gh-134584: Elimiate redundant refcounting from _BINARY_OP_X_UNICODE
corona10 Jun 22, 2025
46b423f
WIP
corona10 Jun 22, 2025
c901079
Address code review
corona10 Jun 22, 2025
ea112a5
Address code review
corona10 Jun 22, 2025
d95aee9
Merge remote-tracking branch 'upstream/main' into gh-134584-unicode
corona10 Jun 23, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion Include/internal/pycore_opcode_metadata.h
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

20 changes: 20 additions & 0 deletions Lib/test/test_capi/test_opt.py
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -5,6 +5,8 @@
import unittest
import gc
import os
import random
import string

import _opcode

Expand DownExpand Up@@ -2362,6 +2364,24 @@ def testfunc(n):
self.assertNotIn("_GUARD_TOS_INT", uops)
self.assertNotIn("_GUARD_NOS_INT", uops)

def test_store_fast_pop_top_specialize_unicode(self):
def random_str(n):
return ''.join(random.choice(string.ascii_uppercase + string.digits) for _ in range(n))
def testfunc(args):
a, b, n = args
c = ''
for _ in range(n):
c += a + b
return c

r0, r1 = random_str(32), random_str(32)
res, ex = self._run_with_optimizer(testfunc, (r0, r1, TIER2_THRESHOLD))
self.assertAlmostEqual(res, TIER2_THRESHOLD * (r0 + r1))
self.assertIsNotNone(ex)
uops = get_opnames(ex)
self.assertIn("_POP_TOP_NOP", uops)
Copy link
Member

@tomasr8tomasr8Jun 22, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe also add

self.assertIn("_BINARY_OP_ADD_UNICODE", uops)

to make sure this is the op that's actually generated?

Should we also test for _POP_TOP_UNICODE?

Copy link
MemberAuthor

@corona10corona10Jun 22, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In this code, It will be replaced to _POP_TOP_NOP because _POP_TOP_UNICODE will be optimized to _POP_TOP_NOP .

Copy link
Member

@tomasr8tomasr8Jun 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yup, do you think it makes sense to test _POP_TOP_UNICODE separately? That is, the case when it is not optimized to _POP_TOP_NOP? (not sure how difficult it'll be to come up with a test case though)

Copy link
Member

@Fidget-SpinnerFidget-SpinnerJun 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have a test for that in my original PR already.

self.assertNotIn("_POP_TOP_UNICODE", uops)

def test_attr_promotion_failure(self):
# We're not testing for any specific uops here, just
# testing it doesn't crash.
Expand Down
4 changes: 2 additions & 2 deletions Lib/test/test_sys.py
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -1671,7 +1671,7 @@ def func():
INTERPRETER_FRAME = '9PihcP'
else:
INTERPRETER_FRAME = '9PhcP'
check(x, size('3PiccPPP' + INTERPRETER_FRAME + 'P'))
check(x, size('3PiccPPP' + INTERPRETER_FRAME + 'PP'))
# function
def func(): pass
check(func, size('16Pi'))
Expand All@@ -1688,7 +1688,7 @@ def bar(cls):
check(bar, size('PP'))
# generator
def get_gen(): yield 1
check(get_gen(), size('6P4c' + INTERPRETER_FRAME + 'P'))
check(get_gen(), size('6P4c' + INTERPRETER_FRAME + 'PP'))
# iterator
check(iter('abc'), size('lP'))
# callable-iterator
Expand Down
2 changes: 1 addition & 1 deletion Programs/test_frozenmain.h
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

8 changes: 4 additions & 4 deletions Python/bytecodes.c
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -772,23 +772,23 @@ dummy_func(
macro(BINARY_OP_SUBTRACT_FLOAT) =
_GUARD_TOS_FLOAT + _GUARD_NOS_FLOAT + unused/5 + _BINARY_OP_SUBTRACT_FLOAT;

pure op(_BINARY_OP_ADD_UNICODE, (left, right -- res)){
pure op(_BINARY_OP_ADD_UNICODE, (left, right -- res, l, r)){
PyObject *left_o = PyStackRef_AsPyObjectBorrow(left);
PyObject *right_o = PyStackRef_AsPyObjectBorrow(right);
assert(PyUnicode_CheckExact(left_o));
assert(PyUnicode_CheckExact(right_o));

STAT_INC(BINARY_OP, hit);
PyObject *res_o = PyUnicode_Concat(left_o, right_o);
PyStackRef_CLOSE_SPECIALIZED(right, _PyUnicode_ExactDealloc);
PyStackRef_CLOSE_SPECIALIZED(left, _PyUnicode_ExactDealloc);
INPUTS_DEAD();
ERROR_IF(res_o == NULL);
l = left;
r = right;
res = PyStackRef_FromPyObjectSteal(res_o);
}

macro(BINARY_OP_ADD_UNICODE) =
_GUARD_TOS_UNICODE + _GUARD_NOS_UNICODE + unused/5 + _BINARY_OP_ADD_UNICODE;
_GUARD_TOS_UNICODE + _GUARD_NOS_UNICODE + unused/5 + _BINARY_OP_ADD_UNICODE + _POP_TOP_UNICODE + _POP_TOP_UNICODE;

// This is a subtle one. It's a super-instruction for
// BINARY_OP_ADD_UNICODE followed by STORE_FAST
Expand Down
4 changes: 4 additions & 0 deletions Python/compile.c
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -1431,6 +1431,10 @@ optimize_and_assemble_code_unit(struct compiler_unit *u, PyObject *const_cache,
&optimized_instrs) < 0){
goto error;
}
/* Reserve an extra word on the stack to ensure there is space for uops to
pass at least one item on the stack to a subsequent uop.
*/
stackdepth++;

/** Assembly **/
co = _PyAssemble_MakeCodeObject(&u->u_metadata, const_cache, consts,
Expand Down
10 changes: 7 additions & 3 deletions Python/executor_cases.c.h
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

18 changes: 16 additions & 2 deletions Python/generated_cases.c.h
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

11 changes: 10 additions & 1 deletion Python/optimizer_bytecodes.c
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -318,7 +318,7 @@ dummy_func(void){
}
}

op(_BINARY_OP_ADD_UNICODE, (left, right -- res)){
op(_BINARY_OP_ADD_UNICODE, (left, right -- res, l, r)){
if (sym_is_const(ctx, left) && sym_is_const(ctx, right)){
assert(PyUnicode_CheckExact(sym_get_const(ctx, left)));
assert(PyUnicode_CheckExact(sym_get_const(ctx, right)));
Expand All@@ -332,6 +332,8 @@ dummy_func(void){
else{
res = sym_new_type(ctx, &PyUnicode_Type);
}
l = left;
r = right;
}

op(_BINARY_OP_INPLACE_ADD_UNICODE, (left, right -- )){
Expand DownExpand Up@@ -561,6 +563,13 @@ dummy_func(void){
value = PyJitRef_Borrow(sym_new_const(ctx, ptr));
}

op(_POP_TOP_UNICODE, (value -- )){
if (PyJitRef_IsBorrowed(value) ||
sym_is_immortal(PyJitRef_Unwrap(value))){
REPLACE_OP(this_instr, _POP_TOP_NOP, 0, 0);
}
}

op(_POP_TOP, (value -- )){
PyTypeObject *typ = sym_get_type(value);
if (PyJitRef_IsBorrowed(value) ||
Expand Down
14 changes: 14 additions & 0 deletions Python/optimizer_cases.c.h
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

Loading