exp / log / pow for complex<float>

xtensor-stack / xsimd

C++ wrappers for SIMD intrinsics and parallelized, optimized mathematical functions (SSE, AVX, AVX512, NEON, SVE))

https://xsimd.readthedocs.io/

BSD 3-Clause "New" or "Revised" License

2.17k stars 254 forks source link

exp / log / pow for complex<float> #139

Open JohanMabille opened 6 years ago

JohanMabille commented 6 years ago

pow of complex<float> has some accuracy issues with AVX512.

mattip commented 5 years ago

Numpy recently merged support for accelerated simd exp/log for float32 in prs numpy/numpy#13581 and others. We seem to have hit an issue numpy/numpy#13586 with clang 7.0, have you experienced anything similar? Also, it seems glibc 2.27 has quite good support out of the box for avx. What do you think?

SylvainCorlay commented 5 years ago

Hey Matti, nice to see you here. Sorry for the late reply. We will definitely look into this numpy issue.