ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-14 02:02:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

experimenting global and buffer load/store

Browse Source
This commit is contained in:
Chao Liu
2019-09-18 02:05:42 -05:00
parent c7a6545ec4
commit 5b7a18c506
4 changed files with 14 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
composable_kernel/include/tensor_operation/threadwise_generic_tensor_slice_copy.hpp
View File

@@ -838,7 +838,7 @@ struct ThreadwiseGenericTensorSliceCopy_v2r1
#if 1 // source code
vector_data = *reinterpret_cast<const src_vector_t*>(
&p_src[src_normal_offset + src_merged_offset]);
#elif 1 // inline asm using global_load
#elif 0 // inline asm using global_load
vector_data = __global_load<TData, SrcDataPerAccess>(
p_src,
static_cast<uint32_t>(src_merged_offset),
@@ -943,7 +943,7 @@ struct ThreadwiseGenericTensorSliceCopy_v2r1
#if 1 // source code
*reinterpret_cast<dst_vector_t*>(
&p_dst[dst_normal_offset + dst_merged_offset]) = vector_data;
#elif 1 // inline asm using global_store
#elif 0 // inline asm using global_store
__global_store<TData, DstDataPerAccess>(
vector_data, p_dst, dst_merged_offset, dst_normal_offset);
#elif 1 // inline asm using buffer_store