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

fix assertion (#1790)

Browse Source
This commit is contained in:
Sean Xiaowen Zhang
2024-09-09 11:05:27 -07:00
committed by GitHub
parent 323c8170bf
commit 21d0534167
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
include/cute/arch/copy_sm80.hpp
View File

@@ -77,7 +77,7 @@ struct SM80_CP_ASYNC_CACHEGLOBAL
using DRegisters = TD[1];
static_assert(sizeof(TS) == sizeof(TD), "cp.async requires sizeof(src_value_type) == sizeof(dst_value_type)");
static_assert(sizeof(TS) == 4 || sizeof(TS) == 8 || sizeof(TS) == 16, "cp.async sizeof(TS) is not supported");
static_assert(sizeof(TS) == 16, "cp.async sizeof(TS) is not supported");
CUTE_HOST_DEVICE static void
copy(TS const& gmem_src,
@@ -134,7 +134,7 @@ struct SM80_CP_ASYNC_CACHEGLOBAL_ZFILL
using DRegisters = TD[1];
static_assert(sizeof(TS) == sizeof(TD), "cp.async requires sizeof(src_value_type) == sizeof(dst_value_type)");
static_assert(sizeof(TS) == 4 || sizeof(TS) == 8 || sizeof(TS) == 16, "cp.async sizeof(TS) is not supported");
static_assert(sizeof(TS) == 16, "cp.async sizeof(TS) is not supported");
CUTE_HOST_DEVICE static void
copy(TS const& gmem_src,