floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #3190 from martin-frbg/issue3128-2 

Replace spurious AVX512 requirement in the Haswell drot microkernel with an AVX2/FMA3 guard
This commit is contained in:
Martin Kroeker 2021-04-27 06:36:28 +02:00 committed by GitHub
parent 0b8a436af9 3d511f0e66
commit adcfe7b789
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
kernel/x86_64/drot_microk_haswell-2.c
View File

@ -1,6 +1,4 @@
/* need a new enough GCC for avx512 support */
#if (( defined(__GNUC__) && __GNUC__ > 6 && defined(__AVX512CD__)) || (defined(__clang__) && __clang_major__ >= 9))
#if defined(HAVE_FMA3) && defined(HAVE_AVX2)
#define HAVE_DROT_KERNEL 1
#include <immintrin.h>