Skip to content

GitLab

"...composable_kernel.git" did not exist on "aebd211c363324ec8be401f17fe815e21da59081"
Commit d3f2dbbd authored by rocking's avatar rocking
Browse files

Change of post shuffle thread descriptor

parent 5490de18
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 2 deletions
include/ck/tensor_operation/gpu/grid/gemm_layernorm/gridwise_gemm_multiple_d_welford_first_half_xdl_cshuffle.hpp
View file @ d3f2dbbd
...@@ -755,7 +755,7 @@ struct GridwiseGemmMultipleDWelfordFirstHalf_xdl_cshuffle ...@@ -755,7 +755,7 @@ struct GridwiseGemmMultipleDWelfordFirstHalf_xdl_cshuffle
// To apply D0, D1, ... and Welford. // To apply D0, D1, ... and Welford.
// threadwise copy from LDS to VGPR // threadwise copy from LDS to VGPR
constexpr auto post_shuffle_thread_cluster_desc = constexpr auto post_shuffle_thread_cluster_desc =
make_cluster_descriptor(PostShuffleThreadClusterSize_M_N{}, Sequence<1, 0>{}); make_cluster_descriptor(PostShuffleThreadClusterSize_M_N{}, Sequence<0, 1>{});
const auto post_shuffle_thread_cluster_idx = const auto post_shuffle_thread_cluster_idx =
post_shuffle_thread_cluster_desc.CalculateBottomIndex( post_shuffle_thread_cluster_desc.CalculateBottomIndex(
...@@ -861,7 +861,7 @@ struct GridwiseGemmMultipleDWelfordFirstHalf_xdl_cshuffle ...@@ -861,7 +861,7 @@ struct GridwiseGemmMultipleDWelfordFirstHalf_xdl_cshuffle
using BlockwiseWelford = BlockwiseWelford<AccDataType, using BlockwiseWelford = BlockwiseWelford<AccDataType,
BlockSize, BlockSize,
PostShuffleThreadClusterSize_M_N, PostShuffleThreadClusterSize_M_N,
Sequence<1, 0>, Sequence<0, 1>,
false>; false>;
constexpr int num_shuffleM = constexpr int num_shuffleM =
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment