8000 BUG: Integer divide tests fail on 32 bit Windows wheel builds. · Issue #20216 · numpy/numpy · GitHub
[go: up one dir, main page]
Skip to content

BUG: Integer divide tests fail on 32 bit Windows wheel builds. #20216

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

Jump to bottom
Closed
charris opened this issue Oct 28, 2021 · 4 comments · Fixed by #20219
Closed

BUG: Integer divide tests fail on 32 bit Windows wheel builds. #20216

charris opened this issue Oct 28, 2021 · 4 comments · Fixed by #20219

Comments

@charris
Copy link
Member
charris commented Oct 28, 2021 •
edited
Loading

This is using vs2019 on Windows server 2019.

FAILED core/tests/test_simd.py::Test_SIMD_ALL_256_AVX2_s64::test_arithmetic_intdiv
FAILED core/tests/test_simd.py::Test_SIMD_ALL_128_SSE42_u64::test_arithmetic_intdiv
FAILED core/tests/test_simd.py::Test_SIMD_ALL_128_SSE42_s64::test_arithmetic_intdiv
FAILED core/tests/test_simd.py::Test_SIMD_ALL_128_baseline_u64::test_arithmetic_intdiv
FAILED core/tests/test_simd.py::Test_SIMD_ALL_128_baseline_s64::test_arithmetic_intdiv
FAILED core/tests/test_umath.py::TestDivision::test_division_int_boundary[int64-np.arange(fo.max-lsize, fo.max, dtype=dtype),np.arange(lsize, dtype=dtype),range(15)]
FAILED core/tests/test_umath.py::TestDivision::test_division_int_boundary[int64-np.arange(fo.min, fo.min+lsize, dtype=dtype),np.arange(lsize//-2, lsize//2, dtype=dtype),range(fo.min, fo.min + 15)]
FAILED core/tests/test_umath.py::TestDivision::test_division_int_boundary[int64-np.arange(fo.max-lsize, fo.max, dtype=dtype),np.arange(lsize, dtype=dtype),[1,3,9,13,neg, fo.min+1, fo.min//2, fo.max//3, fo.max//4]]
FAILED core/tests/test_umath.py::TestDivision::test_division_int_boundary[uint64-np.arange(fo.max-lsize, fo.max, dtype=dtype),np.arange(lsize, dtype=dtype),range(15)]
FAILED core/tests/test_umath.py::TestDivision::test_division_int_boundary[uint64-np.arange(fo.min, fo.min+lsize, dtype=dtype),np.arange(lsize//-2, lsize//2, dtype=dtype),range(fo.min, fo.min + 15)]
FAILED core/tests/test_umath.py::TestDivision::test_division_int_boundary[uint64-np.arange(fo.max-lsize, fo.max, dtype=dtype),np.arange(lsize, dtype=dtype),[1,3,9,13,neg, fo.min+1, fo.min//2, fo.max//3, fo.max//4]]
FAILED ma/tests/test_core.py::TestMaskedArrayInPlaceArithmetic::test_inplace_floor_division_scalar_type
= 16 failed, 18166 passed, 429 skipped, 19 xfailed, 1 xpassed in 2506.14s (0:41:46) =
uh-oh, unmatched shift_free(ptr, 8) but allocated 720
uh-oh, unmatched shift_free(ptr, 8) but allocated 720
uh-oh, unmatched shift_free(ptr, 8) but allocated 720

See https://tinyurl.com/2myhyf8w.
@charris
Copy link
Member Author
charris commented Oct 28, 2021 •
edited
Loading

@mattip The uh-oh, unmatched shift_free(ptr, 8) but allocated 720 seems to be yours.

EDIT: And also shows up on 64 bit windows.
EDIT: And on Linux and Mac

@seberg
Copy link
Member
seberg commented Oct 28, 2021

It seems plausible to me that the test_time_to_time fails for the same reason as the division tests, so would like to wait digging too deep on that until those are cleared out. Or was the situation different (e.g. different machine)?

@charris
Copy link
Member Author
charris commented Oct 28, 2021

@seberg I think you are correct, that crash was a one off.

@mattip
Copy link
Member
mattip commented Oct 28, 2021

The uh-oh, unmatched shift_free(ptr, 8) but allocated 720, as noted in this comment, comes from this pattern np.empty((2, 2, 0, 2, 2)). It calculates the bytes to allocate by multiplying the shape as-if the 0 is 1, but frees 8 bytes.

@seiko2plus seiko2plus mentioned this issue Oct 28, 2021
Merged
@WarrenWeckesser WarrenWeckesser mentioned this issue Nov 3, 2021
Closed
@charris charris mentioned this issue Nov 3, 2021
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

BUG, SIMD: Workaround broadcasting SIMD 64-bit integers on MSVC 32-bit seiko2plus/numpy
3 participants
@seberg @charris @mattip
0