ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 06:49:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK_BUILDER] Convert convolution traits to a struct with factory functions (#3547)

Browse Source 
* Factor helpers out of conv_traits.hpp

* Create a non-templated conv_traits struct

* Migrate to new instance-specific instance_to_conv_traits functions

* Clean up reflection concepts

* Clean up ConvTraits helpers

* Update testing for convolution traits

This is a lot of cleanup on tests to have verbose coverage of feature
extraction, explicit tests for each supported device kernel, and
simple, readable test code.

* Address reviewer comments and resolve merge conflict
This commit is contained in:
John Shumway
2026-01-15 01:03:21 -08:00
committed by GitHub
parent 8705fdcb0c
commit 5122637215
17 changed files with 2288 additions and 1875 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_multiple_abd_xdl_cshuffle.hpp
View File

@@ -2108,7 +2108,7 @@ struct DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle
std::unique_ptr<ck_tile::reflect::Description> describe() const override
{
static_assert(ck_tile::reflect::conv::HasConvTraits<DeviceOp>,
static_assert(ck_tile::reflect::HasConvTraits<DeviceOp>,
"ConvTraits specialization not found for this device operation. "
"If you modified the template parameters of this class, ensure that "
"the corresponding ConvTraits specialization in "

2
include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_multiple_d_xdl_large_tensor_cshuffle.hpp
View File

@@ -1282,7 +1282,7 @@ struct DeviceGroupedConvFwdMultipleD_Xdl_CShuffle_Large_Tensor
std::unique_ptr<ck_tile::reflect::Description> describe() const override
{
static_assert(
ck_tile::reflect::conv::HasConvTraits<DeviceOp>,
ck_tile::reflect::HasConvTraits<DeviceOp>,
"ConvTraits specialization not found for this device operation. "
"If you modified the template parameters of this class, ensure that "
"the corresponding ConvTraits specialization in "