ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-12 17:26:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

refactor

Browse Source
This commit is contained in:
Chao Liu
2019-05-24 16:43:29 -05:00
parent 8fcf3f1e68
commit e17b495dff
5 changed files with 12 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/include/blockwise_merged_tensor_slice_op.hip.hpp → src/include/blockwise_generic_tensor_slice_op.hip.hpp
View File

@@ -14,7 +14,7 @@ template <index_t BlockSize,
class DstAccessOrder,
index_t SrcDataPerRead,
index_t DstDataPerRead>
struct BlockwiseTensorSliceCopy_generic_v1
struct BlockwiseGenericTensorSliceCopy_v1
{
static constexpr index_t nDim = SrcDesc::GetNumOfDimension();
@@ -22,8 +22,8 @@ struct BlockwiseTensorSliceCopy_generic_v1
index_t mDstMyThreadOffset;
__device__
BlockwiseTensorSliceCopy_generic_v1(Array<index_t, nDim> src_block_data_multi_id_begin,
Array<index_t, nDim> dst_block_data_multi_id_begin)
BlockwiseGenericTensorSliceCopy_v1(Array<index_t, nDim> src_block_data_multi_id_begin,
Array<index_t, nDim> dst_block_data_multi_id_begin)
{
// check NDim consistent
static_assert(nDim == SrcDesc::GetNumOfDimension() &&
@@ -155,7 +155,7 @@ struct BlockwiseTensorSliceCopy_generic_v1
const index_t clipboard_offset = thread_tensor_desc.GetOffsetFromMultiIndex(
clipboard_data_multi_id_begin); // cannot not constexpr, why?
threadwise_tensor_slice_copy_generic(SrcDesc{},
threadwise_generic_tensor_slice_copy(SrcDesc{},
p_src + src_offset + mSrcMyThreadOffset,
make_zero_array<index_t, nDim>(),
thread_tensor_desc,
@@ -193,7 +193,7 @@ struct BlockwiseTensorSliceCopy_generic_v1
const index_t dst_offset = DstDesc{}.GetOffsetFromMultiIndex(
dst_data_multi_id_begin); // cannot not constexpr, why?
threadwise_tensor_slice_copy_generic(thread_tensor_desc,
threadwise_generic_tensor_slice_copy(thread_tensor_desc,
p_clipboard + clipboard_offset,
make_zero_array<index_t, nDim>(),
DstDesc{},

2
src/include/gridwise_convolution_implicit_gemm_v1r3_lds_double_buffer_nchw_cyxk_nkhw.hip.hpp
View File

@@ -474,7 +474,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_lds_double_buffer_nchw_cyxk_nkhw
map_out_global2thread,
Number<OutThreadCopyDataPerWrite_W>{});
#else
threadwise_tensor_slice_copy_generic(
threadwise_generic_tensor_slice_copy(
out_10d_thread_desc.ReorderGivenNew2Old(map_out_global2thread),
p_out_thread,
make_zero_array<index_t, 10>(),

2
src/include/gridwise_convolution_implicit_gemm_v1r3_nchw_cyxk_nkhw.hip.hpp
View File

@@ -423,7 +423,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
map_out_global2thread,
Number<OutThreadCopyDataPerWrite_W>{});
#else
threadwise_tensor_slice_copy_generic(
threadwise_generic_tensor_slice_copy(
out_10d_thread_desc.ReorderGivenNew2Old(map_out_global2thread),
p_out_thread,
make_zero_array<index_t, 10>(),

8
src/include/gridwise_convolution_implicit_gemm_v3_nchw_cyxk_nkhw.hip.hpp
View File

@@ -3,7 +3,7 @@
#include "ConstantTensorDescriptor.hip.hpp"
#include "ConstantMergedTensorDescriptor.hip.hpp"
#include "ConstantMatrixDescriptor.hip.hpp"
#include "blockwise_merged_tensor_slice_op.hip.hpp"
#include "blockwise_generic_tensor_slice_op.hip.hpp"
#include "blockwise_gemm.hip.hpp"
#include "threadwise_tensor_slice_op.hip.hpp"
@@ -123,7 +123,7 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw
// input blockwise copy
// slice a merged tensor, reorder and copy to a normal tensor
// this copy operator already has blockwise offset built-in
const auto blockwise_in_copy = BlockwiseTensorSliceCopy_generic_v1<
const auto blockwise_in_copy = BlockwiseGenericTensorSliceCopy_v1<
BlockSize,
Float,
decltype(in_c_n1_b_n2_global_merged_desc),
@@ -152,7 +152,7 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw
// this copy operator already have blockwise offset built-in
const auto blockwise_wei_copy =
#if 0
BlockwiseTensorSliceCopy_generic_v1<BlockSize,
BlockwiseGenericTensorSliceCopy_v1<BlockSize,
Float,
decltype(wei_c_k_global_desc),
decltype(wei_c_k_block_desc),
@@ -318,7 +318,7 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw
out_k_n1_b_n2_global_merged_desc.GetOffsetFromMultiIndex(
k_thread_data_on_global, 0, b_thread_data_on_global, 0);
threadwise_tensor_slice_copy_generic(out_n0_n1_n2_k0_k1_k2_h_w_thread_desc,
threadwise_generic_tensor_slice_copy(out_n0_n1_n2_k0_k1_k2_h_w_thread_desc,
p_out_thread,
{0, 0, 0, 0, 0, 0, 0, 0},
out_n0_n1_n2_k0_k1_k2_h_w_global_mem_desc,

2
src/include/threadwise_tensor_slice_op.hip.hpp
View File

@@ -194,7 +194,7 @@ threadwise_tensor_slice_copy_reorder_given_dst2src_v3(SrcDesc,
}
template <class Float, class SrcDesc, class DstDesc, class SliceLengths, class DimAccessOrder>
__device__ void threadwise_tensor_slice_copy_generic(
__device__ void threadwise_generic_tensor_slice_copy(
SrcDesc,
const Float* __restrict__ p_src,
Array<index_t, SrcDesc::GetNumOfDimension()> src_multi_id_begin,