We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
2 parents 4062c10 + bba75d5 commit 334cd24Copy full SHA for 334cd24
param.h
@@ -3701,6 +3701,12 @@ is a big desktop or server with abundant cache rather than a phone or embedded d
3701
3702
#elif defined(A64FX) // 512-bit SVE
3703
3704
+#if defined(XDOUBLE) || defined(DOUBLE)
3705
+#define GEMM_PREFERED_SIZE 8
3706
+#else
3707
+#define GEMM_PREFERED_SIZE 16
3708
+#endif
3709
+
3710
/* When all BLAS3 routines are implemeted with SVE, SGEMM_DEFAULT_UNROLL_M should be "sve_vl".
3711
Until then, just keep it different than DGEMM_DEFAULT_UNROLL_N to keep copy routines in both directions seperated. */
3712
#define SGEMM_DEFAULT_UNROLL_M 4
0 commit comments