Intel Adds AVX2/FMA Optimized Math Functions To Glibc 2.27
Intel engineers have introduced AVX2/FMA-optimized math functions for glibc and will appear in the project's next stable release.
There is now optimized asin, atan2, exp, expf, log, pow, atan, sin, and tan functions in glibc Git for benefiting from x86-64 for fused multiply-add, as acknowledged by this commit.
For at least some of these math functions, the new optimized versions can be improved by more than 50% on Intel Skylake hardware.
Those not wanting to wait around until the glibc 2.27 release can already find the patches carried by Intel's Clear Linux distribution.
There is now optimized asin, atan2, exp, expf, log, pow, atan, sin, and tan functions in glibc Git for benefiting from x86-64 for fused multiply-add, as acknowledged by this commit.
For at least some of these math functions, the new optimized versions can be improved by more than 50% on Intel Skylake hardware.
Those not wanting to wait around until the glibc 2.27 release can already find the patches carried by Intel's Clear Linux distribution.
28 Comments