ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-12 09:16:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

add back some code

Browse Source
This commit is contained in:
Chao Liu
2019-08-13 12:21:38 -05:00
parent 8bdaba51f8
commit 40836ab926
4 changed files with 54 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
composable_kernel/include/kernel_algorithm/gridwise_convolution_implicit_gemm_v1r3_chwn_cyxk_khwn_lds_double_buffer.hpp
View File

@@ -7,6 +7,10 @@
#include "blockwise_generic_tensor_slice_copy.hpp"
#include "threadwise_generic_tensor_slice_copy.hpp"
#include "blockwise_batched_gemm.hpp"
#include "blockwise_2d_tensor_op.hpp"
#include "blockwise_4d_tensor_op.hpp"
#include "threadwise_tensor_slice_copy.hpp"
#include "threadwise_4d_tensor_op.hpp"
namespace ck {
@@ -129,10 +133,20 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
constexpr auto out_k_h_w_n_thread_desc = make_ConstantTensorDescriptor_packed(
Sequence<KPerThread, HoPerThread, WoPerThread, NPerThread>{});
#if 1
// blockwise copy
// input: format is [C, Hi, Wi, N]
const auto blockwise_in_copy =
Blockwise4dTensorCopy3<BlockSize,
Float,
decltype(in_c_h_w_n_global_desc),
decltype(in_c_h_w_n_block_desc),
decltype(in_c_h_w_n_block_desc.GetLengths()),
InBlockCopyClusterLengths_CHWN,
InBlockCopyDataPerAccess_N>{};
#else
auto blockwise_in_copy =
BlockwiseGenericTensorSliceCopy_v2<BlockSize,
BlockwiseGenericTensorSliceCopy_v1<BlockSize,
decltype(in_c_h_w_n_global_desc),
decltype(in_c_h_w_n_block_desc),
decltype(in_c_h_w_n_block_desc.GetLengths()),
@@ -146,11 +160,21 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
InBlockCopyDataPerAccess_N,
InBlockCopyDataPerAccess_N>({0, 0, 0, 0},
{0, 0, 0, 0});
#endif
#if 1
// blockwise wei copy
// format is [CPerBlock, X * KPerBlock]
const auto blockwise_wei_copy =
BlockwiseGenericTensorSliceCopy_v2<BlockSize,
Blockwise2dTensorCopy3<BlockSize,
Float,
decltype(wei_c_k_global_desc),
decltype(wei_c_k_block_desc),
decltype(wei_c_k_block_desc.GetLengths()),
WeiBlockCopyDataPerAccess_K>({0, 0}, {0, 0});
#else
const auto blockwise_wei_copy =
BlockwiseGenericTensorSliceCopy_v1<BlockSize,
decltype(wei_c_k_global_desc),
decltype(wei_c_k_block_desc),
decltype(wei_c_k_block_desc.GetLengths()),
@@ -163,6 +187,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
1,
WeiBlockCopyDataPerAccess_K,
WeiBlockCopyDataPerAccess_K>({0, 0}, {0, 0});
#endif
// a series of blockwise batched GEMM
// C_matrix += transpose(A_matrix) * B_matrix
@@ -402,7 +427,15 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
wo_block_data_begin + wo_thread_data_begin,
n_block_data_begin + n_thread_data_begin);
ThreadwiseGenericTensorSliceCopy_v2r1<decltype(out_10d_thread_desc),
#if 1
threadwise_tensor_slice_copy(out_10d_thread_desc,
p_out_thread,
out_10d_global_desc,
p_out_thread_on_global,
out_10d_thread_desc.GetLengths(),
Number<OutThreadCopyDataPerAccess_N>{});
#else
ThreadwiseGenericTensorSliceCopy_v1r1<decltype(out_10d_thread_desc),
decltype(out_10d_global_desc),
decltype(out_10d_thread_desc.GetLengths()),
arithmetic_sequence_gen<0, 10, 1>::type,
@@ -413,6 +446,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
OutThreadCopyDataPerAccess_N>(
make_zero_array<index_t, 10>(), make_zero_array<index_t, 10>())
.Run(p_out_thread, p_out_thread_on_global);
#endif
}).Else([&](auto fwd) {
static_assert(fwd(GemmNPerThreadSubC) >= NPerBlock && NPerThread == NPerBlock &&
GemmNPerThreadSubC % NPerThread == 0,
@@ -460,7 +494,15 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
wo_block_data_begin + wo_thread_data_begin,
n_block_data_begin + n_thread_data_begin);
ThreadwiseGenericTensorSliceCopy_v2r1<decltype(out_10d_thread_desc),
#if 1
threadwise_tensor_slice_copy(out_10d_thread_desc,
p_out_thread,
out_10d_global_desc,
p_out_thread_on_global,
out_10d_thread_desc.GetLengths(),
Number<OutThreadCopyDataPerAccess_N>{});
#else
ThreadwiseGenericTensorSliceCopy_v1r1<decltype(out_10d_thread_desc),
decltype(out_10d_global_desc),
decltype(out_10d_thread_desc.GetLengths()),
arithmetic_sequence_gen<0, 10, 1>::type,
@@ -471,6 +513,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn_lds_double_buffer
OutThreadCopyDataPerAccess_N>(
make_zero_array<index_t, 10>(), make_zero_array<index_t, 10>())
.Run(p_out_thread, p_out_thread_on_global);
#endif
});
}
};

4
composable_kernel/include/tensor_description/tensor_coordinate.hpp
View File

@@ -301,14 +301,14 @@ struct TensorCoordinate
private:
template <class... Ts>
__host__ __device__ static constexpr auto
MakeDummyTensorCoordinate(ConstantTensorDescriptor<Ts...>)
MakeDummyTensorCoordinate(ConstantTensorDescriptor<Ts...>)
{
return NormalTensorCoordinate<ConstantTensorDescriptor<Ts...>>();
}
template <class... Ts>
__host__ __device__ static constexpr auto
MakeDummyTensorCoordinate(ConstantMergedTensorDescriptor<Ts...>)
MakeDummyTensorCoordinate(ConstantMergedTensorDescriptor<Ts...>)
{
return MergedTensorCoordinate<ConstantMergedTensorDescriptor<Ts...>>();
}

5
driver/include/device_convolution_implicit_gemm_v1_chwn_cyxk_khwn.hpp
View File

@@ -143,7 +143,7 @@ void device_convolution_implicit_gemm_v1_chwn_cyxk_khwn(InDesc,
constexpr index_t WeiBlockCopyDataPerAccess_K = 4;
constexpr index_t OutThreadCopyDataPerAccess_N = 2;
#elif 1
#elif 0
// for 3x3, 34x34, v1r3, Pascal
// for 3x3, 28x28, v1r3, Pascal
// for 3x3, 14x14, v1r3, Pascal
@@ -266,9 +266,12 @@ void device_convolution_implicit_gemm_v1_chwn_cyxk_khwn(InDesc,
constexpr index_t GemmDataPerReadA = 4;
constexpr index_t GemmDataPerReadB = 4;
using InBlockCopySubLengths_CHWN = Sequence<1, 1, 1, 4>;
using InBlockCopyClusterLengths_CHWN = Sequence<8, 2, 4, 4>;
constexpr index_t InBlockCopyDataPerAccess_N = 4;
using WeiBlockCopySubLengths_CK = Sequence<1, 4>;
using WeiBlockCopyClusterLengths_CK = Sequence<8, 32>;
constexpr index_t WeiBlockCopyDataPerAccess_K = 4;
constexpr index_t OutThreadCopyDataPerAccess_N = 4;

2
driver/src/driver.cpp
View File

@@ -71,7 +71,7 @@ int main(int argc, char* argv[])
{
using namespace ck;
#if 1
#if 0
constexpr index_t N = 64;
constexpr index_t C = 1536;
constexpr index_t HI = 8;