Skip to content

GitLab

Commit baaad9ec authored by Astha Rai's avatar Astha Rai
Browse files

updated vector dim access to enable vector load

parent f3b6e205
Hide whitespace changes
Inline Side-by-side
Showing with 4 additions and 4 deletions
example/44_elementwise_permute/elementwise_permute_3d_ndhwc.cpp
View file @ baaad9ec
......@@ -28,8 +28,8 @@ using DeviceElementwisePermuteInstance =
8, // MPerThread
8, // NPerThread
8, // KPerThread
ck::Sequence<1>, // InScalarPerVectorSeq
ck::Sequence<1>>; // OutScalarPerVectorSeq
ck::Sequence<8>, // InScalarPerVectorSeq
ck::Sequence<4>>; // OutScalarPerVectorSeq
template <typename HostTensorA, typename HostTensorB, typename Functor>
void host_elementwise4D(HostTensorB& B_ndhwc, const HostTensorA& A_ncdhw, Functor functor)
......
include/ck/tensor_operation/gpu/grid/gridwise_elementwise_3d.hpp
View file @ baaad9ec
......@@ -142,7 +142,7 @@ struct GridwiseElementwise_3D
decltype(thread_buffer_desc_mnk),
Sequence<MPerThread, NPerThread, KPerThread>, // SliceLengths
Sequence<0, 1, 2>, // DimAccessOrder
0, // SrcVectorDim
01, // SrcVectorDim
InScalarPerVectorSeq::At(I), // InScalarPerVectorSeq::At(I), //
// ScalarPerVector
1, // SrcScalarStrideInVector
......@@ -163,7 +163,7 @@ struct GridwiseElementwise_3D
PassThroughOp,
Sequence<MPerThread, NPerThread, KPerThread>, // SliceLengths
Sequence<0, 1, 2>, // DimAccessOrder
1, // SrcVectorDim
2, // SrcVectorDim
OutScalarPerVectorSeq::At(I), // OutScalarPerVectorSeq::At(I),
InMemoryDataOperationEnum::Set,
1,
......
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