amd/blis
1
0
Fork 0
mirror of https://github.com/amd/blis.git synced 2026-05-11 17:50:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

Minor fixes for 8x24 KNL kernel.

Browse Source
This commit is contained in:
Devin Matthews
2016-07-25 10:02:25 -05:00
parent 65735bbedf
commit e0cce177cc
1 changed files with 7 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
kernels/x86_64/knl/3/bli_dgemm_opt_8x24.c
View File

@@ -38,11 +38,11 @@
extern int32_t offsets[24];
#define A_PREFETCH_DIST 5
#define A_PREFETCH_DIST 10
#define PREFETCH_A 1
#define PIPELINE_A 1
#define UNROLL_X2 0
#define UNROLL_X4 1
#define UNROLL_X2 1
#define UNROLL_X4 0
#define UPDATE_SCATTERED(n) \
KMOV(K(1), ESI) \
@@ -609,8 +609,7 @@ void bli_dgemm_opt_8x24
MOV(RDI, VAR(cs_c))
LEA(RDI, MEM(,RDI,8))
MOV(R8, MEM(RBX))
MOV(RDX, VAR(rs_c))
VBROADCASTSS(YMM(5), MEM(RDX))
VBROADCASTSS(YMM(4), VAR(rs_c))
//MOV(RAX, 0xCC)
//MOV(RBX, 0xF0)
//MOV(RSI, 0xAA)
@@ -623,11 +622,11 @@ void bli_dgemm_opt_8x24
//VPADDD(ZMM(4) MASK_K(1), ZMM(4), ZMM(2))
//VPADDD(ZMM(4) MASK_K(2), ZMM(4), ZMM(3))
MOV(RSI, VAR(offsetPtr))
VMOVAPS(YMM(5), MEM(RSI))
VMOVUPS(YMM(5), MEM(RSI))
VPMULLD(YMM(4), YMM(5), YMM(4))
MOV(RSI, 0xFF)
MOV(RSI, IMM(0xFF))
SAL1(R8) // shift out the sign bit to check for +/- zero
//JZ(.DGENSTORBZ)
JZ(.DGENSTORBZ)
UPDATE_SCATTERED( 8)
UPDATE_SCATTERED( 9)