ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-03 21:21:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

add conv1d/3d bwd weight instances (#318)

Browse Source 
* add conv1d/3d bwd weight instances

* add profiler code
This commit is contained in:
Shaojie WANG
2022-07-09 04:42:20 +08:00
committed by GitHub
parent 4fe9c393b8
commit 763ca61581
19 changed files with 1907 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

129
include/ck/tensor_operation/gpu/device/device_convnd_backward_weight_xdl_c_shuffle_nhwc_kyxc_nhwk.hpp
View File

@@ -996,22 +996,46 @@ struct DeviceConvndBwdWeightXdl_C_Shuffle_Input_N_Hi_Wi_C_Weight_K_Y_X_C_Output_
0,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_.GetElementSpaceSize() *
sizeof(CDataType)));
float elapsed_time =
launch_and_time_kernel(stream_config,
kernel,
dim3(grid_size),
dim3(BlockSize),
0,
arg.p_a_grid_,
arg.p_b_grid_,
arg.p_c_grid_,
arg.a_grid_desc_kbatch_k0_m_k1_,
arg.b_grid_desc_kbatch_k0_n_k1_,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_,
arg.a_element_op_,
arg.b_element_op_,
arg.c_element_op_,
arg.block_2_ctile_map_);
return launch_and_time_kernel(stream_config,
kernel,
dim3(grid_size),
dim3(BlockSize),
0,
arg.p_a_grid_,
arg.p_b_grid_,
arg.p_c_grid_,
arg.a_grid_desc_kbatch_k0_m_k1_,
arg.b_grid_desc_kbatch_k0_n_k1_,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_,
arg.a_element_op_,
arg.b_element_op_,
arg.c_element_op_,
arg.block_2_ctile_map_);
hipGetErrorString(hipMemset(
arg.p_c_grid_,
0,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_.GetElementSpaceSize() *
sizeof(CDataType)));
launch_and_time_kernel(StreamConfig{nullptr, false},
kernel,
dim3(grid_size),
dim3(BlockSize),
0,
arg.p_a_grid_,
arg.p_b_grid_,
arg.p_c_grid_,
arg.a_grid_desc_kbatch_k0_m_k1_,
arg.b_grid_desc_kbatch_k0_n_k1_,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_,
arg.a_element_op_,
arg.b_element_op_,
arg.c_element_op_,
arg.block_2_ctile_map_);
return elapsed_time;
};
// run kernel for bf16 with splitk
@@ -1022,21 +1046,46 @@ struct DeviceConvndBwdWeightXdl_C_Shuffle_Input_N_Hi_Wi_C_Weight_K_Y_X_C_Output_
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_.GetElementSpaceSize() *
sizeof(AccDataType)));
return launch_and_time_kernel(stream_config,
kernel,
dim3(grid_size),
dim3(BlockSize),
0,
arg.p_a_grid_,
arg.p_b_grid_,
static_cast<AccDataType*>(arg.p_workspace_),
arg.a_grid_desc_kbatch_k0_m_k1_,
arg.b_grid_desc_kbatch_k0_n_k1_,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_,
arg.a_element_op_,
arg.b_element_op_,
arg.c_element_op_,
arg.block_2_ctile_map_);
float elapsed_time =
launch_and_time_kernel(stream_config,
kernel,
dim3(grid_size),
dim3(BlockSize),
0,
arg.p_a_grid_,
arg.p_b_grid_,
static_cast<AccDataType*>(arg.p_workspace_),
arg.a_grid_desc_kbatch_k0_m_k1_,
arg.b_grid_desc_kbatch_k0_n_k1_,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_,
arg.a_element_op_,
arg.b_element_op_,
arg.c_element_op_,
arg.block_2_ctile_map_);
hipGetErrorString(hipMemset(
arg.p_workspace_,
0,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_.GetElementSpaceSize() *
sizeof(AccDataType)));
launch_and_time_kernel(StreamConfig{nullptr, false},
kernel,
dim3(grid_size),
dim3(BlockSize),
0,
arg.p_a_grid_,
arg.p_b_grid_,
static_cast<AccDataType*>(arg.p_workspace_),
arg.a_grid_desc_kbatch_k0_m_k1_,
arg.b_grid_desc_kbatch_k0_n_k1_,
arg.c_grid_desc_mblock_mperblock_nblock_nperblock_,
arg.a_element_op_,
arg.b_element_op_,
arg.c_element_op_,
arg.block_2_ctile_map_);
return elapsed_time;
};
// kernel for type conversion
@@ -1210,6 +1259,20 @@ struct DeviceConvndBwdWeightXdl_C_Shuffle_Input_N_Hi_Wi_C_Weight_K_Y_X_C_Output_
static bool IsSupportedArgument(const Argument& arg)
{
if constexpr(ConvBackwardWeightSpecialization ==
ConvolutionBackwardWeightSpecialization::Filter1x1Stride1Pad0)
{
// check if it's 1x1, stride=1 pad = 0 conv
for(int i = 0; i < NumDimSpatial; i++)
{
if(!(arg.filter_spatial_lengths_[i] == 1 && arg.conv_filter_strides_[i] == 1 &&
arg.input_left_pads_[i] == 0 && arg.input_right_pads_[i] == 0))
{
return false;
}
}
}
// vector load A/B matrix from global memory
if(!(ABlockTransferSrcVectorDim == 2 && BBlockTransferSrcVectorDim == 2 &&
arg.Conv_K_ % ABlockTransferSrcScalarPerVector == 0 &&
@@ -1334,6 +1397,12 @@ struct DeviceConvndBwdWeightXdl_C_Shuffle_Input_N_Hi_Wi_C_Weight_K_Y_X_C_Output_
<< NPerBlock << ", "
<< K0PerBlock
<< ">";
if constexpr(ConvBackwardWeightSpecialization ==
ConvolutionBackwardWeightSpecialization::Filter1x1Stride1Pad0){
str << " Filter1x1Stride1Pad0";
}
// clang-format on
return str.str();