Skip to content

Optimize unpack, str.__add__ and fastlocals #7293

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

Merged
youknowone merged 9 commits into from
Mar 2, 2026
Merged

Optimize unpack, str.__add__ and fastlocals #7293

Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
ef3c5ef
Remove intermediate Vec allocation in unpack_sequence fast path
youknowone Mar 1, 2026
b6760e1
Use read-only atomic load before swap in check_signals
youknowone Mar 1, 2026
98ea46e
Cache builtins downcast in ExecutingFrame for LOAD_GLOBAL
youknowone Mar 1, 2026
f3a8d8d
Add number Add slot to PyStr for direct str+str dispatch
youknowone Mar 1, 2026
67ed603
Guard FastLocals access in locals() with try_lock on state mutex
youknowone Mar 1, 2026
b1cf47c
Use exact type check for builtins_dict cache
youknowone Mar 1, 2026
0af466c
Consolidate with_recursion in _cmp to single guard
youknowone Mar 1, 2026
5b9e078
Add opcode-level fast paths for FOR_ITER, COMPARE_OP, BINARY_OP
youknowone Mar 1, 2026
4c24868
Also check globals is exact dict for LOAD_GLOBAL fast path
youknowone Mar 1, 2026
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
Prev Previous commit
Next Next commit
Consolidate with_recursion in _cmp to single guard 
Move the recursion depth check to wrap the entire _cmp body
instead of each individual call_cmp direction, reducing Cell
read/write pairs and scopeguard overhead per comparison.
  • Loading branch information
@youknowone
youknowone committed Mar 1, 2026
commit 0af466c24cd45910bac471e10123a18861ad82c2
21 changes: 15 additions & 6 deletions crates/vm/src/protocol/object.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -277,13 +277,24 @@ impl PyObject {

// Perform a comparison, raising TypeError when the requested comparison
// operator is not supported.
// see: CPython PyObject_RichCompare
// see: CPython PyObject_RichCompare / do_richcompare
#[inline] // called by ExecutingFrame::execute_compare with const op
fn _cmp(
&self,
other: &Self,
op: PyComparisonOp,
vm: &VirtualMachine,
) -> PyResult<Either<PyObjectRef, bool>> {
// Single recursion guard for the entire comparison (matching CPython's
// Py_EnterRecursiveCallTstate placement in do_richcompare).
vm.with_recursion("in comparison", || self._cmp_inner(other, op, vm))
}

fn _cmp_inner(
&self,
other: &Self,
op: PyComparisonOp,
vm: &VirtualMachine,
) -> PyResult<Either<PyObjectRef, bool>> {
let swapped = op.swapped();
let call_cmp = |obj: &Self, other: &Self, op| {
Expand All @@ -302,19 +313,17 @@ impl PyObject {
!self_class.is(other_class) && other_class.fast_issubclass(self_class)
};
if is_strict_subclass {
let res = vm.with_recursion("in comparison", || call_cmp(other, self, swapped))?;
let res = call_cmp(other, self, swapped)?;
checked_reverse_op = true;
if let PyArithmeticValue::Implemented(x) = res {
return Ok(x);
}
}
if let PyArithmeticValue::Implemented(x) =
vm.with_recursion("in comparison", || call_cmp(self, other, op))?
{
if let PyArithmeticValue::Implemented(x) = call_cmp(self, other, op)? {
return Ok(x);
}
if !checked_reverse_op {
let res = vm.with_recursion("in comparison", || call_cmp(other, self, swapped))?;
let res = call_cmp(other, self, swapped)?;
if let PyArithmeticValue::Implemented(x) = res {
return Ok(x);
}
Expand Down