ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-15 18:42:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

tidy yp

Browse Source 
[ROCm/composable_kernel commit: 471830a052]
This commit is contained in:
Chao Liu
2019-04-09 18:07:36 -05:00
parent 4fba215ede
commit 75ca00f748
7 changed files with 98 additions and 111 deletions
Download Patch File Download Diff File Expand all files Collapse all files

62
driver/device_implicit_gemm_convolution_1_chwn_cyxk_khwn.hpp
View File

@@ -346,37 +346,37 @@ void device_implicit_gemm_convolution_1_chwn_cyxk_khwn(InDesc,
#elif 0
GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
#endif
<GridSize,
BlockSize,
T,
decltype(in_chwn_desc),
decltype(wei_cyxk_desc),
decltype(out_khwn_desc),
NPerBlock,
KPerBlock,
CPerBlock,
HoPerBlock,
WoPerBlock,
NPerThread,
KPerThread,
HoPerThread,
WoPerThread,
GemmMPerThreadSubC,
GemmNPerThreadSubC,
GemmMLevel0Cluster,
GemmNLevel0Cluster,
GemmMLevel1Cluster,
GemmNLevel1Cluster,
GemmKPerThreadLoop,
GemmDataPerReadA,
GemmDataPerReadB,
Sequence<InBlockCopy_ThreadPerDimC,
InBlockCopy_ThreadPerDimH,
InBlockCopy_ThreadPerDimW,
InBlockCopy_ThreadPerDimN>,
InBlockCopyDataPerRead,
WeiBlockCopyDataPerRead,
OutThreadCopyDataPerWrite>{};
<GridSize,
BlockSize,
T,
decltype(in_chwn_desc),
decltype(wei_cyxk_desc),
decltype(out_khwn_desc),
NPerBlock,
KPerBlock,
CPerBlock,
HoPerBlock,
WoPerBlock,
NPerThread,
KPerThread,
HoPerThread,
WoPerThread,
GemmMPerThreadSubC,
GemmNPerThreadSubC,
GemmMLevel0Cluster,
GemmNLevel0Cluster,
GemmMLevel1Cluster,
GemmNLevel1Cluster,
GemmKPerThreadLoop,
GemmDataPerReadA,
GemmDataPerReadB,
Sequence<InBlockCopy_ThreadPerDimC,
InBlockCopy_ThreadPerDimH,
InBlockCopy_ThreadPerDimW,
InBlockCopy_ThreadPerDimN>,
InBlockCopyDataPerRead,
WeiBlockCopyDataPerRead,
OutThreadCopyDataPerWrite>{};
float time = launch_kernel(run_gridwise_convolution<decltype(gridwise_conv), T>,
dim3(GridSize),

6
src/include/ConstantTensorDescriptor.hip.hpp
View File

@@ -381,7 +381,8 @@ __host__ __device__ void print_ConstantTensorDescriptor(TDesc, const char* s)
constexpr auto I7 = Number<7>{};
constexpr auto I8 = Number<8>{};
printf("%s dim %u, lengths {%u %u %u %u %u %u %u %u %u}, strides {%u %u %u %u %u %u %u %u %u}\n",
printf("%s dim %u, lengths {%u %u %u %u %u %u %u %u %u}, strides {%u %u %u %u %u %u %u %u "
"%u}\n",
s,
desc.GetDimension(),
desc.GetLength(I0),
@@ -416,7 +417,8 @@ __host__ __device__ void print_ConstantTensorDescriptor(TDesc, const char* s)
constexpr auto I8 = Number<8>{};
constexpr auto I9 = Number<9>{};
printf("%s dim %u, lengths {%u %u %u %u %u %u %u %u %u %u}, strides {%u %u %u %u %u %u %u %u %u %u}\n",
printf("%s dim %u, lengths {%u %u %u %u %u %u %u %u %u %u}, strides {%u %u %u %u %u %u %u "
"%u %u %u}\n",
s,
desc.GetDimension(),
desc.GetLength(I0),

11
src/include/blockwise_4d_tensor_op.hip.hpp
View File

@@ -577,8 +577,8 @@ struct Blockwise4dTensorCopy3
iloop_d3 * thread_per_d3 * DataPerRead);
*(reinterpret_cast<vector_t*>(&p_dst[dst_offset + mDstMyThreadOffset])) =
*(reinterpret_cast<const vector_t*>(&p_src[src_offset +
mSrcMyThreadOffset]));
*(reinterpret_cast<const vector_t*>(
&p_src[src_offset + mSrcMyThreadOffset]));
}
}
}
@@ -612,7 +612,8 @@ struct Blockwise4dTensorCopy3
return DataPerRead * nloop_d0 * nloop_d1 * nloop_d2 * nloop_d3;
}
__device__ void RunLoadRegisterClipboard(const Float* __restrict__ p_src, Float* __restrict__ p_clipboard) const
__device__ void RunLoadRegisterClipboard(const Float* __restrict__ p_src,
Float* __restrict__ p_clipboard) const
{
constexpr auto I0 = Number<0>{};
constexpr auto I1 = Number<1>{};
@@ -670,8 +671,8 @@ struct Blockwise4dTensorCopy3
iloop_d3 * thread_per_d3 * DataPerRead);
*(reinterpret_cast<vector_t*>(&p_clipboard[dst_offset])) =
*(reinterpret_cast<const vector_t*>(&p_src[src_offset +
mSrcMyThreadOffset]));
*(reinterpret_cast<const vector_t*>(
&p_src[src_offset + mSrcMyThreadOffset]));
}
}
}

26
src/include/gridwise_convolution_implicit_gemm_v1_chwn_cyxk_khwn.hip.hpp
View File

@@ -43,8 +43,9 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn
Float* const __restrict__ p_out_global) const
{
// be careful of this assertion
static_assert(NPerThread <= NPerBlock && NPerBlock % NPerThread == 0,
"wrong! should satisfy: NPerThread <= NPerBlock && NPerBlock % NPerThread == 0");
static_assert(
NPerThread <= NPerBlock && NPerBlock % NPerThread == 0,
"wrong! should satisfy: NPerThread <= NPerBlock && NPerBlock % NPerThread == 0");
constexpr auto I0 = Number<0>{};
constexpr auto I1 = Number<1>{};
@@ -69,8 +70,9 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn
constexpr index_t WiPerBlock = WoPerBlock + X - 1;
// divide block work: [K, Ho, Wo, N]
static_assert(N % NPerBlock == 0 && K % KPerBlock == 0 && C % CPerBlock == 0 && Ho % HoPerBlock == 0 && Wo % WoPerBlock == 0,
"wrong! cannot evenly divide work for workgroup ");
static_assert(N % NPerBlock == 0 && K % KPerBlock == 0 && C % CPerBlock == 0 &&
Ho % HoPerBlock == 0 && Wo % WoPerBlock == 0,
"wrong! cannot evenly divide work for workgroup ");
constexpr index_t KBlockWork = (K + KPerBlock - 1) / KPerBlock;
constexpr index_t HBlockWork = (Ho + HoPerBlock - 1) / HoPerBlock;
@@ -101,8 +103,7 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn
mod_conv::max(index_t(4), InBlockCopyDataPerRead, WeiBlockCopyDataPerRead);
constexpr auto in_chwn_block_desc = make_ConstantTensorDescriptor_aligned(
Sequence<CPerBlock, HiPerBlock, WiPerBlock, NPerBlock>{},
Number<max_align>{});
Sequence<CPerBlock, HiPerBlock, WiPerBlock, NPerBlock>{}, Number<max_align>{});
constexpr auto wei_ek_block_desc = make_ConstantTensorDescriptor_aligned(
Sequence<CPerBlock * Y * X, KPerBlock>{}, Number<max_align>{});
@@ -280,17 +281,8 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn
constexpr index_t K2 = GemmMPerThreadSubC;
constexpr index_t K1 = KPerBlock / KPerThread;
constexpr auto out_10d_global_desc =
make_ConstantTensorDescriptor(Sequence<K / (K1 * K2),
K1,
K2,
Ho,
Wo / (W1 * W2),
W1,
W2,
N / (N1 * N2),
N1,
N2>{});
constexpr auto out_10d_global_desc = make_ConstantTensorDescriptor(
Sequence<K / (K1 * K2), K1, K2, Ho, Wo / (W1 * W2), W1, W2, N / (N1 * N2), N1, N2>{});
constexpr auto out_10d_thread_desc = make_ConstantTensorDescriptor(
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, 1, 1, N2>{});

43
src/include/gridwise_convolution_implicit_gemm_v1_chwn_cyxk_khwn_lds_double_buffer.hip.hpp
View File

@@ -43,8 +43,9 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
Float* const __restrict__ p_out_global) const
{
// be careful of this assertion
static_assert(NPerThread <= NPerBlock && NPerBlock % NPerThread == 0,
"wrong! should satisfy: NPerThread <= NPerBlock && NPerBlock % NPerThread == 0");
static_assert(
NPerThread <= NPerBlock && NPerBlock % NPerThread == 0,
"wrong! should satisfy: NPerThread <= NPerBlock && NPerBlock % NPerThread == 0");
constexpr auto I0 = Number<0>{};
constexpr auto I1 = Number<1>{};
@@ -72,8 +73,9 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
static_assert(C % (2 * CPerBlock) == 0, "C cannot be evenly divided");
// divide block work: [K, Ho, Wo, N]
static_assert(N % NPerBlock == 0 && K % KPerBlock == 0 && C % CPerBlock == 0 && Ho % HoPerBlock == 0 && Wo % WoPerBlock == 0,
"wrong! cannot evenly divide work for workgroup ");
static_assert(N % NPerBlock == 0 && K % KPerBlock == 0 && C % CPerBlock == 0 &&
Ho % HoPerBlock == 0 && Wo % WoPerBlock == 0,
"wrong! cannot evenly divide work for workgroup ");
constexpr index_t KBlockWork = (K + KPerBlock - 1) / KPerBlock;
constexpr index_t HBlockWork = (Ho + HoPerBlock - 1) / HoPerBlock;
@@ -104,8 +106,7 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
mod_conv::max(index_t(4), InBlockCopyDataPerRead, WeiBlockCopyDataPerRead);
constexpr auto in_chwn_block_desc = make_ConstantTensorDescriptor_aligned(
Sequence<CPerBlock, HiPerBlock, WiPerBlock, NPerBlock>{},
Number<max_align>{});
Sequence<CPerBlock, HiPerBlock, WiPerBlock, NPerBlock>{}, Number<max_align>{});
constexpr auto wei_ek_block_desc = make_ConstantTensorDescriptor_aligned(
Sequence<CPerBlock * Y * X, KPerBlock>{}, Number<max_align>{});
@@ -250,16 +251,15 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
blockwise_wei_copy.RunLoadRegisterClipboard(p_wei_global_block_offset,
p_wei_register_clipboard);
// a series of batched GEMM
for(index_t y = 0; y < Y; ++y)
{
for(index_t x = 0; x < X; ++x)
{
blockwise_batch_gemm.Run(p_wei_block_now +
wei_cyxk_block_desc.Get1dIndex(0, y, x, 0),
p_in_block_now + in_chwn_block_desc.Get1dIndex(0, y, x, 0),
p_out_thread);
blockwise_batch_gemm.Run(
p_wei_block_now + wei_cyxk_block_desc.Get1dIndex(0, y, x, 0),
p_in_block_now + in_chwn_block_desc.Get1dIndex(0, y, x, 0),
p_out_thread);
}
}
@@ -291,10 +291,10 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
{
for(index_t x = 0; x < X; ++x)
{
blockwise_batch_gemm.Run(p_wei_block_double +
wei_cyxk_block_desc.Get1dIndex(0, y, x, 0),
p_in_block_double + in_chwn_block_desc.Get1dIndex(0, y, x, 0),
p_out_thread);
blockwise_batch_gemm.Run(
p_wei_block_double + wei_cyxk_block_desc.Get1dIndex(0, y, x, 0),
p_in_block_double + in_chwn_block_desc.Get1dIndex(0, y, x, 0),
p_out_thread);
}
}
@@ -376,17 +376,8 @@ struct GridwiseConvolutionImplicitGemm_v1_chwn_cyxk_khwn_lds_double_buffer
constexpr index_t K2 = GemmMPerThreadSubC;
constexpr index_t K1 = KPerBlock / KPerThread;
constexpr auto out_10d_global_desc =
make_ConstantTensorDescriptor(Sequence<K / (K1 * K2),
K1,
K2,
Ho,
Wo / (W1 * W2),
W1,
W2,
N / (N1 * N2),
N1,
N2>{});
constexpr auto out_10d_global_desc = make_ConstantTensorDescriptor(
Sequence<K / (K1 * K2), K1, K2, Ho, Wo / (W1 * W2), W1, W2, N / (N1 * N2), N1, N2>{});
constexpr auto out_10d_thread_desc = make_ConstantTensorDescriptor(
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, 1, 1, N2>{});

52
src/include/gridwise_convolution_implicit_gemm_v1r2_chwn_cyxk_khwn.hip.hpp
View File

@@ -43,8 +43,9 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
Float* const __restrict__ p_out_global) const
{
// be careful of this assertion
static_assert(NPerThread <= NPerBlock && NPerBlock % NPerThread == 0,
"wrong! should satisfy: NPerThread <= NPerBlock && NPerBlock % NPerThread == 0");
static_assert(
NPerThread <= NPerBlock && NPerBlock % NPerThread == 0,
"wrong! should satisfy: NPerThread <= NPerBlock && NPerBlock % NPerThread == 0");
constexpr auto I0 = Number<0>{};
constexpr auto I1 = Number<1>{};
@@ -69,8 +70,9 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
constexpr index_t WiPerBlock = WoPerBlock + X - 1;
// divide block work: [K, Ho, Wo, N]
static_assert(N % NPerBlock == 0 && K % KPerBlock == 0 && C % CPerBlock == 0 && Ho % HoPerBlock == 0 && Wo % WoPerBlock == 0,
"wrong! cannot evenly divide work for workgroup ");
static_assert(N % NPerBlock == 0 && K % KPerBlock == 0 && C % CPerBlock == 0 &&
Ho % HoPerBlock == 0 && Wo % WoPerBlock == 0,
"wrong! cannot evenly divide work for workgroup ");
constexpr index_t KBlockWork = (K + KPerBlock - 1) / KPerBlock;
constexpr index_t HBlockWork = (Ho + HoPerBlock - 1) / HoPerBlock;
@@ -93,7 +95,8 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
const index_t wi_block_data_begin = wo_block_data_begin;
// 2d tensor view of gridwise weight
constexpr auto wei_ck_global_desc = make_ConstantTensorDescriptor(Sequence<C, K>{}, Sequence<Y * X * K, 1>{});
constexpr auto wei_ck_global_desc =
make_ConstantTensorDescriptor(Sequence<C, K>{}, Sequence<Y * X * K, 1>{});
// tensor view of blockwise input and weight in LDS
// be careful of alignment
@@ -124,7 +127,7 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
// blockwise wei copy
// format is [CPerBlock, KPerBlock]
const auto blockwise_wei_copy =
#if 0//debug
#if 0 // debug
Blockwise2dTensorCopy1<BlockSize,
Float,
decltype(wei_ck_global_desc),
@@ -139,14 +142,16 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
WeiBlockCopyDataPerRead>{};
#endif
// a series of blockwise batched GEMM
// C_matrix += transpose(A_matrix) * B_matrix
// A_matrix and B_matrix saved in LDS, C_matrix saved in register
// A_matrix[C,K] is a sub-matrix of wei_block[C,K]
// B_matrix[C,Wo*N] is a sub-matrix of in_block[C,Hi,Wi,N]
// C_matrix[K,Wo*N] is a sub-matrix of out_block[K,Ho,Wo,N]
constexpr auto a_cxk_block_mtx_desc = make_ConstantMatrixDescriptor(
Number<CPerBlock>{}, Number<KPerBlock>{}, Number<wei_ck_block_desc.GetStride(I0)>{});
// a series of blockwise batched GEMM
// C_matrix += transpose(A_matrix) * B_matrix
// A_matrix and B_matrix saved in LDS, C_matrix saved in register
// A_matrix[C,K] is a sub-matrix of wei_block[C,K]
// B_matrix[C,Wo*N] is a sub-matrix of in_block[C,Hi,Wi,N]
// C_matrix[K,Wo*N] is a sub-matrix of out_block[K,Ho,Wo,N]
constexpr auto a_cxk_block_mtx_desc =
make_ConstantMatrixDescriptor(Number<CPerBlock>{},
Number<KPerBlock>{},
Number<wei_ck_block_desc.GetStride(I0)>{});
constexpr auto b_cxwn_block_mtx_desc =
make_ConstantMatrixDescriptor(Number<CPerBlock>{},
@@ -180,7 +185,7 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
GemmDataPerReadB>{};
// LDS: be careful of alignment
constexpr index_t in_block_space = in_chwn_block_desc.GetElementSpace(Number<max_align>{});
constexpr index_t in_block_space = in_chwn_block_desc.GetElementSpace(Number<max_align>{});
constexpr index_t wei_block_space = wei_ck_block_desc.GetElementSpace(Number<max_align>{});
__shared__ Float p_in_block[in_block_space];
@@ -227,8 +232,8 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
{
// weight: global mem to LDS
blockwise_wei_copy.Run(p_wei_global_block_offset +
wei_cyxk_global_desc.Get1dIndex(0, y, x, 0),
p_wei_block);
wei_cyxk_global_desc.Get1dIndex(0, y, x, 0),
p_wei_block);
__syncthreads();
@@ -297,17 +302,8 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
constexpr index_t K2 = GemmMPerThreadSubC;
constexpr index_t K1 = KPerBlock / KPerThread;
constexpr auto out_10d_global_desc =
make_ConstantTensorDescriptor(Sequence<K / (K1 * K2),
K1,
K2,
Ho,
Wo / (W1 * W2),
W1,
W2,
N / (N1 * N2),
N1,
N2>{});
constexpr auto out_10d_global_desc = make_ConstantTensorDescriptor(
Sequence<K / (K1 * K2), K1, K2, Ho, Wo / (W1 * W2), W1, W2, N / (N1 * N2), N1, N2>{});
constexpr auto out_10d_thread_desc = make_ConstantTensorDescriptor(
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, 1, 1, N2>{});

9
src/include/threadwise_gemm.hip.hpp
View File

@@ -1,6 +1,11 @@
#pragma once
template <class Float, class SrcMatrix, class DstMatrix, index_t NRow, index_t NCol, index_t DataPerRead>
template <class Float,
class SrcMatrix,
class DstMatrix,
index_t NRow,
index_t NCol,
index_t DataPerRead>
__device__ void threadwise_matrix_copy(SrcMatrix,
const Float* __restrict__ p_src,
DstMatrix,
@@ -22,7 +27,7 @@ __device__ void threadwise_matrix_copy(SrcMatrix,
const index_t src_index = src_mtx.Get1dIndex(i, j);
const index_t dst_index = dst_mtx.Get1dIndex(i, j);
*reinterpret_cast<vector_t*>(&p_dst[dst_index]) =
*reinterpret_cast<vector_t*>(&p_dst[dst_index]) =
*reinterpret_cast<const vector_t*>(&p_src[src_index]);
}
}