Skip to content

Commit 430c11e

Browse files
authored
Add files via upload
1 parent fbacd26 commit 430c11e

File tree

1 file changed

+5
-3
lines changed

1 file changed

+5
-3
lines changed

kernel/x86_64/sgemm_kernel_16x4_skylakex_2.c

Lines changed: 5 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -54,15 +54,17 @@
5454
#define COMPUTE_m16(ndim) \
5555
INIT_m16n##ndim\
5656
"movq %%r13,%4; movq %%r14,%1; leaq (%1,%%r12,2),%%r15; addq %%r12,%%r15; movq %2,%5;"\
57-
"cmpq $16,%4; jb "#ndim"016162f;"\
57+
"cmpq $18,%4; jb "#ndim"016162f;"\
5858
#ndim"016161:\n\t"\
5959
KERNEL_k1m16n##ndim\
6060
KERNEL_k1m16n##ndim\
61+
KERNEL_k1m16n##ndim\
6162
"prefetcht1 (%5); prefetcht1 63(%5); addq %3,%5;"\
6263
KERNEL_k1m16n##ndim\
6364
KERNEL_k1m16n##ndim\
64-
"prefetcht1 (%8); addq $"#ndim",%8;"\
65-
"subq $4,%4; cmpq $16,%4; jnb "#ndim"016161b;"\
65+
KERNEL_k1m16n##ndim\
66+
"prefetcht1 (%8); addq $32,%8;"\
67+
"subq $6,%4; cmpq $18,%4; jnb "#ndim"016161b;"\
6668
"movq %2,%5;"\
6769
#ndim"016162:\n\t"\
6870
"testq %4,%4; jz "#ndim"016163f;"\

0 commit comments

Comments
 (0)