done: explicitly separate offset component into compile-time, block-invariant and per-thread components. Experimenting

This commit is contained in:
Chao Liu
2019-09-22 03:17:41 -05:00
parent 51884fc214
commit 6c2c50b020
8 changed files with 351 additions and 202 deletions

View File

@@ -22,7 +22,7 @@
#include "amd_inline_asm.hpp"
#endif
#if CK_USE_AMD_INTRINCIS
#if CK_USE_AMD_INTRINSIC
#include "amd_intrinsic.hpp"
#endif