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

fix typo (#1279)

Browse Source
This commit is contained in:
Jee Li 2024-01-05 01:41:39 +08:00 committed by GitHub
parent 5c756eb774
commit c9591a694d
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
examples/12_gemm_bias_relu/gemm_bias_relu.cu
View File

@ -81,7 +81,7 @@ using ShapeMMAThreadBlock =
// This code section describes tile size a warp will compute // This code section describes tile size a warp will compute
using ShapeMMAWarp = cutlass::gemm::GemmShape<64, 64, 32>; // <- warp tile M = 64, N = 64, K = 32 using ShapeMMAWarp = cutlass::gemm::GemmShape<64, 64, 32>; // <- warp tile M = 64, N = 64, K = 32
// This code section describes the size of MMA op // This code section describes the size of MMA op
using ShapeMMAOp = cutlass::gemm::GemmShape<16, 8, 8>; // <- MMA Op tile M = 8, N = 8, K = 4 using ShapeMMAOp = cutlass::gemm::GemmShape<16, 8, 8>; // <- MMA Op tile M = 16, N = 8, K = 8
// This code section describes how threadblocks are scheduled on GPU // This code section describes how threadblocks are scheduled on GPU
using SwizzleThreadBlock = cutlass::gemm::threadblock::GemmIdentityThreadblockSwizzle<>; // <- ?? using SwizzleThreadBlock = cutlass::gemm::threadblock::GemmIdentityThreadblockSwizzle<>; // <- ??