Skip to content

Commit 9579bd4

Browse files
committed
Modifying a couple paramaters in the "POWER10"-specific section of param.h, for performance enhancements for SGEMM and DGEMM.
1 parent ef0238b commit 9579bd4

File tree

1 file changed

+4
-4
lines changed

1 file changed

+4
-4
lines changed

param.h

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -2455,13 +2455,13 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
24552455
#define ZGEMM_DEFAULT_UNROLL_M 8
24562456
#define ZGEMM_DEFAULT_UNROLL_N 2
24572457

2458-
#define SGEMM_DEFAULT_P 832
2459-
#define DGEMM_DEFAULT_P 320
2458+
#define SGEMM_DEFAULT_P 512
2459+
#define DGEMM_DEFAULT_P 384
24602460
#define CGEMM_DEFAULT_P 512
24612461
#define ZGEMM_DEFAULT_P 256
24622462

2463-
#define SGEMM_DEFAULT_Q 1026
2464-
#define DGEMM_DEFAULT_Q 960
2463+
#define SGEMM_DEFAULT_Q 512
2464+
#define DGEMM_DEFAULT_Q 512
24652465
#define CGEMM_DEFAULT_Q 1026
24662466
#define ZGEMM_DEFAULT_Q 1026
24672467

0 commit comments

Comments
 (0)