leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Fix several typos (#1169)

Browse Source 
Co-authored-by: isaacw <isaacw@nvidia.com>
This commit is contained in:
wang-y-z
2023-11-03 11:54:46 +08:00
committed by GitHub
parent c008b4aea8
commit 557be3ab0e
21 changed files with 30 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/cutlass/transform/collective/sm90_wgmma_transpose.hpp
View File

@@ -556,7 +556,7 @@ public:
constexpr auto WarpThreadLayout = make_layout(make_shape(Int<WarpThreadShapeN>{}, Int<WarpThreadShapeK>{}));
//////////////////////////////////////////////////////////////////////////////////////////////////////////////
/// A warp group uses 8 steps to transpose the whole WarpgroupTileSize x WarpgroupTileSize.
/// Divide a warp_group_tile into 8x8 warp_tiles to futher reduce the reg usage.
/// Divide a warp_group_tile into 8x8 warp_tiles to further reduce the reg usage.
/// Step 0: Step 1: Step 2: Step 3:
/// W0 W1 W2 W3 -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
/// W1 W0 -- -- -- -- -- -- -- -- W3 W2 -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --