Skip to content

Commit 934e601

Browse files
authored
Update dgemm_kernel_4x8_skylakex_2.c
1 parent eb1e9c8 commit 934e601

File tree

1 file changed

+5
-3
lines changed

1 file changed

+5
-3
lines changed

kernel/x86_64/dgemm_kernel_4x8_skylakex_2.c

Lines changed: 5 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -97,15 +97,17 @@
9797

9898
/* %10 for prefetch of C elements before storage; %4 = ldc(in bytes),%11 for prefetch of next B block */
9999
#define INNER_KERNELm8(nn) \
100-
"movq %3,%10;cmpq $16,%2;jb "#nn"001f;"\
100+
"movq %3,%10;cmpq $18,%2;jb "#nn"001f;"\
101101
#nn"008:\n\t"\
102102
INNER_KERNEL_k1m8n##nn "addq $64,%1;"\
103103
INNER_KERNEL_k1m8n##nn "addq $64,%1;"\
104+
INNER_KERNEL_k1m8n##nn "addq $64,%1;"\
104105
"prefetcht1 (%10); prefetcht1 63(%10); addq %4,%10;"\
105106
INNER_KERNEL_k1m8n##nn "addq $64,%1;"\
106107
INNER_KERNEL_k1m8n##nn "addq $64,%1;"\
107-
"prefetcht1 (%11); addq $16,%11;"\
108-
"subq $4,%2;cmpq $16,%2;jnb "#nn"008b;"\
108+
INNER_KERNEL_k1m8n##nn "addq $64,%1;"\
109+
"prefetcht1 (%11); addq $32,%11;"\
110+
"subq $6,%2;cmpq $18,%2;jnb "#nn"008b;"\
109111
"movq %3,%10;"\
110112
#nn"001:\n\t"\
111113
"cmpq $1,%2;jb "#nn"000f;"\

0 commit comments

Comments
 (0)