🌐 AI搜索 & 代理 主页
Skip to content

BUG: fix matmul with transposed out arg #29179

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
seberg merged 3 commits into from
Jun 12, 2025
Merged

BUG: fix matmul with transposed out arg #29179

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2ebabb7
BUG: fix matmul with transposed out arg
mattip Jun 11, 2025
ece0bdb
DOC: add release note
mattip Jun 11, 2025
6100fba
fixes from review
mattip Jun 12, 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
Prev Previous commit
fixes from review
  • Loading branch information
@mattip
mattip committed Jun 12, 2025
commit 6100fba37b5f3c9584437d91d5c8c3c219021622
6 changes: 0 additions & 6 deletions doc/release/upcoming_changes/23752.performance.rst
View file
Open in desktop

This file was deleted.

4 changes: 4 additions & 0 deletions doc/release/upcoming_changes/29179.change.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
Fix bug in ``matmul`` for non-contiguous out kwarg parameter
------------------------------------------------------------
In some cases, if ``out`` was non-contiguous, ``np.matmul`` would cause
memory corruption or a c-level assert. This was new to v2.3.0 and fixed in v2.3.1.
6 changes: 6 additions & 0 deletions doc/source/release/2.3.0-notes.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -414,6 +414,12 @@ the best performance.

(`gh-28769 <https://github.com/numpy/numpy/pull/28769>`__)

Performance improvements for ``np.matmul``
------------------------------------------
Enable using BLAS for ``matmul`` even when operands are non-contiguous by copying
if needed.

(`gh-23752 <https://github.com/numpy/numpy/pull/23752>`__)

Changes
=======
Expand Down
10 changes: 6 additions & 4 deletions numpy/_core/tests/test_multiarray.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7272,10 +7272,6 @@ def test_out_contiguous(self):
assert_array_equal(c, tgt_mv)
c = self.matmul(v, a.T, out=out[:, 0, 0])
assert_array_equal(c, tgt_mv)
# issue 29164
out_f = np.zeros((10, 4), dtype=float)
c = self.matmul(a, b, out=out_f[::-2, ::-2])
assert_array_equal(c, tgt)

# test out contiguous in only last dim
out = np.ones((10, 2), dtype=float)
Expand Down Expand Up @@ -7321,6 +7317,12 @@ def test_dot_equivalent(self, args):
r3 = np.matmul(args[0].copy(), args[1].copy())
assert_equal(r1, r3)

# matrix matrix, issue 29164
if [len(args[0].shape), len(args[1].shape)] == [2, 2]:
out_f = np.zeros((r2.shape[0] * 2, r2.shape[1] * 2), order='F')
r4 = np.matmul(*args, out=out_f[::2, ::2])
assert_equal(r2, r4)

def test_matmul_object(self):
import fractions

Expand Down
Loading