BUG: invalid value encountered in expm1 when AVX512 enabled #22772
Description
Describe the issue:
Invalid value encountered in expm1 if NaN set when SVML/AVX512 enabled and only double precision affected
Reproduce the code example:
sde -- python
Python 3.10.8 (main, Nov 1 2022, 14:18:21) [GCC 12.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> np.seterr(all='raise')
{'divide': 'warn', 'over': 'warn', 'under': 'ignore', 'invalid': 'warn'}
>>> np.lib.utils._opt_info()
'SSE SSE2 SSE3 SSSE3* SSE41* POPCNT* SSE42* AVX* F16C* FMA3* AVX2* AVX512F* AVX512CD* AVX512_SKX* AVX512_CLX* AVX512_CNL* AVX512_ICL*'
>>> np.expm1(np.array([np.nan], dtype='d'))
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
FloatingPointError: invalid value encountered in expm1Error message:
FloatingPointError: invalid value encountered in expm1NumPy/Python version information:
1.25.0.dev0+167.g4ca8204c5 3.10.8 (main, Nov 1 2022, 14:18:21) [GCC 12.2.0]
Context for the issue:
No response