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

Merge pull request #3669 from VFerrari/fix_small_matrix_kernel 

POWER: fix issues with the small matrix kernel
This commit is contained in:
Martin Kroeker 2022-06-28 16:09:36 +02:00 committed by GitHub
parent 92275a7902 2062280c6f
commit be5500e704
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Makefile.system
View File

@ -261,7 +261,7 @@ endif
#For small matrix optimization #For small matrix optimization
ifeq ($(ARCH), x86_64) ifeq ($(ARCH), x86_64)
SMALL_MATRIX_OPT = 1 SMALL_MATRIX_OPT = 1
else ifeq ($(CORE), POWER10) else ifeq ($(ARCH), power)
SMALL_MATRIX_OPT = 1 SMALL_MATRIX_OPT = 1
endif endif
ifeq ($(SMALL_MATRIX_OPT), 1) ifeq ($(SMALL_MATRIX_OPT), 1)

4
kernel/power/gemm_small_kernel_permit_power10.c
View File

@ -69,6 +69,7 @@ int CNAME(int transa, int transb, BLASLONG M, BLASLONG N, BLASLONG K, FLOAT alph
#endif #endif
#ifdef SMP
// Multi-threading execution outperforms (or approaches) the execution of the // Multi-threading execution outperforms (or approaches) the execution of the
// small kernel. // small kernel.
if (num_cpu_avail(3) > 1) { if (num_cpu_avail(3) > 1) {
@ -77,6 +78,9 @@ int CNAME(int transa, int transb, BLASLONG M, BLASLONG N, BLASLONG K, FLOAT alph
} else { } else {
return 1; return 1;
} }
#else
return 1;
#endif
#endif #endif