Skip to content

[CK] Fix example_grouped_conv_bwd_data_xdl_fp16 with ksplit = 2 #2943

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
illsilin merged 1 commit into from
Sep 29, 2025
Merged

[CK] Fix example_grouped_conv_bwd_data_xdl_fp16 with ksplit = 2 #2943

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 8 additions & 5 deletions ...sor_operation/gpu/device/impl/device_grouped_conv_bwd_data_multiple_d_xdl_cshuffle_v1.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1670,7 +1670,10 @@ struct DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
valid = false;
}
}
else
}
else
{
if constexpr(NXdlPerWave32 > 0)
{
if(!GridwiseGemmCTranspose32::CheckValidity(
arg.a_grid_desc_m_k_container_[i],
Expand All @@ -1685,10 +1688,10 @@ struct DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
valid = false;
}
}
if(!valid)
{
return false;
}
}
if(!valid)
{
return false;
}
}

Expand Down
9 changes: 5 additions & 4 deletions include/ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_xdl_cshuffle.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -561,9 +561,10 @@ struct GridwiseGemmMultipleD_xdl_cshuffle
return;
}

const index_t num_k_per_block =
const index_t num_ak0_per_block =
__builtin_amdgcn_readfirstlane(a_grid_desc_ak0_m_ak1.GetLength(I0) / k_batch);

const index_t num_bk0_per_block =
__builtin_amdgcn_readfirstlane(b_grid_desc_bk0_n_bk1.GetLength(I0) / k_batch);
// HACK: this force m/n_block_data_idx_on_grid into SGPR
const index_t m_block_data_idx_on_grid =
__builtin_amdgcn_readfirstlane(block_work_idx[I0] * MPerBlock);
Expand Down Expand Up @@ -605,7 +606,7 @@ struct GridwiseGemmMultipleD_xdl_cshuffle
true,
NumGemmKPrefetchStage>(
a_grid_desc_ak0_m_ak1,
make_multi_index(num_k_per_block * k_idx, m_block_data_idx_on_grid, 0),
make_multi_index(num_ak0_per_block * k_idx, m_block_data_idx_on_grid, 0),
a_element_op,
a_block_desc_ak0_m_ak1,
make_multi_index(0, 0, 0),
Expand Down Expand Up @@ -636,7 +637,7 @@ struct GridwiseGemmMultipleD_xdl_cshuffle
true,
NumGemmKPrefetchStage>(
b_grid_desc_bk0_n_bk1,
make_multi_index(num_k_per_block * k_idx, n_block_data_idx_on_grid, 0),
make_multi_index(num_bk0_per_block * k_idx, n_block_data_idx_on_grid, 0),
b_element_op,
b_block_desc_bk0_n_bk1,
make_multi_index(0, 0, 0),
Expand Down
Loading