Skip to content

BUG: fix absolute() returning -NaN for +NaN on scalar path #269

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
ndgrigorian merged 3 commits into from
Sep 12, 2026
Merged

BUG: fix absolute() returning -NaN for +NaN on scalar path #269

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
ffa8135
BUG: fix absolute() returning -NaN for +NaN on scalar path
vchamarthi Sep 11, 2026
78501a3
remove unecessary comments
vchamarthi Sep 11, 2026
316e22f
add tests and update changelog with pr
vchamarthi Sep 11, 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
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
* Pinned Cython in the Coverity Scan workflow so generated code stays stable between scans, and added `coverity/README.md` documenting the known Cython-boilerplate false positives and the scan review checklist [gh-266](https://github.com/IntelPython/mkl_umath/pull/266)

### Fixed
* Fixed `absolute` (float32/float64) returning `-NaN` for `+NaN` input on the scalar fallback path (same issue as numpy/numpy@dc478c58b9, gh-31433) [gh-269](https://github.com/IntelPython/mkl_umath/pull/269)
* Fixed an over-decref of the borrowed module dictionary reference on the module initialization error path [gh-264](https://github.com/IntelPython/mkl_umath/pull/264)
* Fixed a leak of the module object when the NumPy C-API import fails during module initialization [gh-263](https://github.com/IntelPython/mkl_umath/pull/263)

Expand Down
2 changes: 1 addition & 1 deletion mkl_umath/src/mkl_umath_loops.c.src
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -917,7 +917,7 @@ mkl_umath_@TYPE@_absolute(char **args, const npy_intp *dimensions, const npy_int
else {
UNARY_LOOP {
const @type@ in1 = *(@type@ *)ip1;
const @type@ tmp = in1 > 0 ? in1 : -in1;
const @type@ tmp = fabs@c@(in1);
/* add 0 to clear -0.0 */
*((@type@ *)op1) = tmp + 0;
}
Expand Down
13 changes: 13 additions & 0 deletions mkl_umath/tests/test_basic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -189,3 +189,16 @@ def test_reduce_complex(func, dtype):
assert np.allclose(
mkl_res, np_res
), f"Results for '{func}[reduce]' do not match"


@pytest.mark.parametrize("size", [100, 8192 + 1])
@pytest.mark.parametrize("dtype", [np.float32, np.float64])
def test_absolute_nan_signbit(size, dtype):
# size 100 takes the fallback loop, 8193 takes MKL's vAbs
for nan in (np.nan, -np.nan):
a = np.full(size, nan, dtype=dtype)
sign = "-" if np.signbit(nan) else "+"
assert not np.signbit(mu.absolute(a)).any(), (
f"absolute({sign}nan) must be +nan for "
f"{dtype.__name__} at size {size}"
)
Loading