amd/blis
1
0
Fork 0
mirror of https://github.com/amd/blis.git synced 2026-04-20 15:48:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

Changing ZGEMM SUP threshold logic for zen5 to fix performance regression (#233)

Browse Source 
- Revert the logical operator from OR (||) to AND (&&) in the DCOMPLEX
  (ZGEMM) SUP threshold condition for k <= 128. The previous change to
  OR logic was causing performance regressions for certain input sizes
  by incorrectly routing cases to the SUP path when the native path
  would be more optimal.
This commit is contained in:
S, Hari Govind
2025-10-17 11:25:33 +05:30
committed by GitHub
parent 1dca574a9d
commit 0ce45e3147
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
kernels/zen5/aocl_smart/bli_aocl_smart_zen5.c
View File

@@ -87,7 +87,7 @@ bool bli_cntx_gemmsup_thresh_is_met_zen5( obj_t* a, obj_t* b, obj_t* c, cntx_t*
// The threshold for m is a single value, but for n, it is
// also based on the packing size of A, since the kernels are
// column preferential
if( ( ( m <= 1380 ) || ( n <= 1520 ) || ( k <= 128 ) ) && ( m + n + k < 6400 ) ) return TRUE;
if ( ( ( m <= 1380 ) || ( n <= 1520 ) && ( k <= 128 ) ) && ( m + n + k < 6400 ) ) return TRUE;
return FALSE;
}