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

Update dgemm_kernel_16x2_skylakex.c

This commit is contained in:
wjc404 2020-02-06 02:14:10 +00:00 committed by GitHub
parent 8b5cdcc64c
commit 3447d04eaf
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
kernel/x86_64/dgemm_kernel_16x2_skylakex.c
View File

@ -393,9 +393,10 @@
#define INIT_m1n12 INIT_m1n10 "vpxor %%xmm9,%%xmm9,%%xmm9;"
#define COMPUTE_SIMPLE(mdim,ndim)\
init_update_k(mdim) INIT_m##mdim##n##ndim "testq %%r13,%%r13; jz 7"#mdim"7"#ndim"9f;"\
init_update_k(mdim) INIT_m##mdim##n##ndim\
"movq %%r13,%5;" INIT_set_papb(mdim,ndim)\
kernel_kstart_n##ndim(mdim,subq)\
"testq %5,%5; jz 7"#mdim"7"#ndim"9f;"\
"7"#mdim"7"#ndim"1:\n\t"\
KERNEL_k1m##mdim##n##ndim "decq %5; jnz 7"#mdim"7"#ndim"1b;"\
"7"#mdim"7"#ndim"9:\n\t"\