mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-13 01:36:06 +00:00
added implicit gemm v1r3, refactored decomposition of wei tensor (loop over y, x first, and C second) to allow easy lds double buffer on C
This commit is contained in:
Chao Liu
@@ -36,7 +36,7 @@ template <index_t GridSize,
|
||||
index_t InBlockCopyDataPerRead,
|
||||
index_t WeiBlockCopyDataPerRead,
|
||||
index_t OutThreadCopyDataPerWrite>
|
||||
struct GridwiseConvolutionImplicitGemm_v1r1_chwn_cyxk_khwn_lds_double_buffer
|
||||
struct GridwiseConvolutionImplicitGemm_v1r1_lds_double_buffer_chwn_cyxk_khwn
|
||||
{
|
||||
__device__ void Run(const Float* const __restrict__ p_in_global,
|
||||
const Float* const __restrict__ p_wei_global,
|
||||
@@ -213,6 +213,7 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
|
||||
// set threadwise output tensor to 0
|
||||
threadwise_4d_tensor_set_zero(out_k_h_w_n_thread_desc, p_out_thread);
|
||||
|
||||
#if 1
|
||||
const Float* p_in_global_block_offset =
|
||||
p_in_global + in_c_h_w_n_global_desc.Get1dIndex(
|
||||
0, hi_block_data_begin, wi_block_data_begin, n_block_data_begin);
|
||||
@@ -247,6 +248,43 @@ struct GridwiseConvolutionImplicitGemm_v1r2_chwn_cyxk_khwn
|
||||
__syncthreads();
|
||||
}
|
||||
}
|
||||
#else
|
||||
// this use much more register, haven't figure out why?
|
||||
for(index_t y = 0; y < Y; ++y)
|
||||
{
|
||||
const Float* p_in_global_block_offset =
|
||||
p_in_global +
|
||||
in_c_h_w_n_global_desc.Get1dIndex(
|
||||
0, hi_block_data_begin + y, wi_block_data_begin, n_block_data_begin);
|
||||
|
||||
const Float* p_wei_global_block_offset =
|
||||
p_wei_global + wei_c_y_x_k_global_desc.Get1dIndex(0, y, 0, k_block_data_begin);
|
||||
|
||||
for(index_t c_block_data_begin = 0; c_block_data_begin < C;
|
||||
c_block_data_begin += CPerBlock,
|
||||
p_in_global_block_offset +=
|
||||
CPerBlock * in_c_h_w_n_global_desc.GetStride(I0),
|
||||
p_wei_global_block_offset +=
|
||||
CPerBlock * wei_c_y_x_k_global_desc.GetStride(I0))
|
||||
{
|
||||
blockwise_in_copy.Run(p_in_global_block_offset, p_in_block);
|
||||
|
||||
blockwise_wei_copy.Run(p_wei_global_block_offset, p_wei_block);
|
||||
|
||||
__syncthreads();
|
||||
|
||||
for(index_t x = 0; x < X; ++x)
|
||||
{
|
||||
blockwise_batch_gemm.Run(p_wei_block + wei_c_x_k_block_desc.Get1dIndex(0, x, 0),
|
||||
p_in_block +
|
||||
in_c_h_w_n_block_desc.Get1dIndex(0, 0, x, 0),
|
||||
p_out_thread);
|
||||
}
|
||||
|
||||
__syncthreads();
|
||||
}
|
||||
}
|
||||
#endif
|
||||
|
||||
// output: register to global mem,
|
||||
const auto c_thread_mtx_begin =
|
||||
|
||||
@@ -35,10 +35,10 @@ template <index_t GridSize,
|
||||
index_t GemmDataPerReadB,
|
||||
class InBlockReorderSrcSubLengths_NCHW,
|
||||
class InBlockReorderSrcClusterLengths_NCHW,
|
||||
class InBlockReorderMapThreadCluster2SrcCluster,
|
||||
class InBlockReorderMapThreadCluster2SrcCluster_CHNW2NCHW,
|
||||
index_t InBlockReorderDataPerRead_W,
|
||||
index_t InBlockReorderDataPerWrite_N,
|
||||
class WeiBlockCopyClusterLengths_KXC,
|
||||
class WeiBlockCopyClusterLengths_CXK,
|
||||
index_t WeiBlockCopyDataPerRead_C,
|
||||
index_t OutThreadCopyDataPerWrite_N>
|
||||
struct GridwiseConvolutionImplicitGemm_v1r2_nchw_cyxk_khwn
|
||||
@@ -122,7 +122,8 @@ struct GridwiseConvolutionImplicitGemm_v1r2_nchw_cyxk_khwn
|
||||
|
||||
// blockwise copy
|
||||
// input: format is [N, C, Hi, Wi] to [C, Hi, Wi, N]
|
||||
auto map_chwn2nchw = Sequence<1, 2, 3, 0>{};
|
||||
constexpr auto map_chwn2nchw = Sequence<1, 2, 3, 0>{};
|
||||
|
||||
const auto blockwise_in_copy_reorder =
|
||||
Blockwise4dTensorCopyReorder3<BlockSize,
|
||||
Float,
|
||||
@@ -132,7 +133,7 @@ struct GridwiseConvolutionImplicitGemm_v1r2_nchw_cyxk_khwn
|
||||
InBlockReorderSrcSubLengths_NCHW,
|
||||
InBlockReorderSrcClusterLengths_NCHW,
|
||||
decltype(map_chwn2nchw),
|
||||
InBlockReorderMapThreadCluster2SrcCluster,
|
||||
InBlockReorderMapThreadCluster2SrcCluster_CHNW2NCHW,
|
||||
InBlockReorderDataPerRead_W,
|
||||
InBlockReorderDataPerWrite_N>{};
|
||||
|
||||
@@ -144,7 +145,7 @@ struct GridwiseConvolutionImplicitGemm_v1r2_nchw_cyxk_khwn
|
||||
decltype(wei_c_x_k_global_desc),
|
||||
decltype(wei_c_x_k_block_desc),
|
||||
decltype(wei_c_x_k_block_desc.GetLengths()),
|
||||
Sequence<4, 1, 32>,
|
||||
WeiBlockCopyClusterLengths_CXK,
|
||||
WeiBlockCopyDataPerRead_C>{};
|
||||
|
||||
// a series of blockwise batched GEMM
|
||||
|
||||
@@ -0,0 +1,331 @@
|
||||
#pragma once
|
||||
#include "common.hip.hpp"
|
||||
#include "ConstantTensorDescriptor.hip.hpp"
|
||||
#include "ConstantMatrixDescriptor.hip.hpp"
|
||||
#include "blockwise_2d_tensor_op.hip.hpp"
|
||||
#include "blockwise_4d_tensor_op.hip.hpp"
|
||||
#include "threadwise_nd_tensor_op.hip.hpp"
|
||||
#include "threadwise_4d_tensor_op.hip.hpp"
|
||||
#include "blockwise_batched_gemm.hip.hpp"
|
||||
|
||||
template <index_t GridSize,
|
||||
index_t BlockSize,
|
||||
class Float,
|
||||
class InGlobalDesc,
|
||||
class WeiGlobalDesc,
|
||||
class OutGlobalDesc,
|
||||
index_t NPerBlock,
|
||||
index_t KPerBlock,
|
||||
index_t CPerBlock,
|
||||
index_t HoPerBlock,
|
||||
index_t WoPerBlock,
|
||||
index_t NPerThread,
|
||||
index_t KPerThread,
|
||||
index_t HoPerThread,
|
||||
index_t WoPerThread,
|
||||
index_t GemmMPerThreadSubC,
|
||||
index_t GemmNPerThreadSubC,
|
||||
index_t GemmMLevel0Cluster,
|
||||
index_t GemmNLevel0Cluster,
|
||||
index_t GemmMLevel1Cluster,
|
||||
index_t GemmNLevel1Cluster,
|
||||
index_t GemmKPerThreadLoop,
|
||||
index_t GemmDataPerReadA,
|
||||
index_t GemmDataPerReadB,
|
||||
class InBlockCopyClusterLengths_CHWN,
|
||||
index_t InBlockCopyDataPerRead_N,
|
||||
index_t WeiBlockCopyDataPerRead_K,
|
||||
index_t OutThreadCopyDataPerWrite_N>
|
||||
struct GridwiseConvolutionImplicitGemm_v1r3_chwn_cyxk_khwn
|
||||
{
|
||||
__device__ void Run(const Float* const __restrict__ p_in_global,
|
||||
const Float* const __restrict__ p_wei_global,
|
||||
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");
|
||||
|
||||
constexpr auto I0 = Number<0>{};
|
||||
constexpr auto I1 = Number<1>{};
|
||||
constexpr auto I2 = Number<2>{};
|
||||
constexpr auto I3 = Number<3>{};
|
||||
|
||||
constexpr auto in_c_h_w_n_global_desc = InGlobalDesc{};
|
||||
constexpr auto wei_c_y_x_k_global_desc = WeiGlobalDesc{};
|
||||
constexpr auto out_k_h_w_n_global_desc = OutGlobalDesc{};
|
||||
|
||||
constexpr index_t C = in_c_h_w_n_global_desc.GetLength(I0);
|
||||
|
||||
constexpr index_t K = out_k_h_w_n_global_desc.GetLength(I0);
|
||||
constexpr index_t Ho = out_k_h_w_n_global_desc.GetLength(I1);
|
||||
constexpr index_t Wo = out_k_h_w_n_global_desc.GetLength(I2);
|
||||
constexpr index_t N = out_k_h_w_n_global_desc.GetLength(I3);
|
||||
|
||||
constexpr index_t Y = wei_c_y_x_k_global_desc.GetLength(I1);
|
||||
constexpr index_t X = wei_c_y_x_k_global_desc.GetLength(I2);
|
||||
|
||||
constexpr index_t HiPerBlock = HoPerBlock + Y - 1;
|
||||
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 ");
|
||||
|
||||
constexpr index_t KBlockWork = (K + KPerBlock - 1) / KPerBlock;
|
||||
constexpr index_t HBlockWork = (Ho + HoPerBlock - 1) / HoPerBlock;
|
||||
constexpr index_t WBlockWork = (Wo + WoPerBlock - 1) / WoPerBlock;
|
||||
constexpr index_t NBlockWork = (N + NPerBlock - 1) / NPerBlock;
|
||||
|
||||
const index_t k_block_work_id = get_block_1d_id() / (HBlockWork * WBlockWork * NBlockWork);
|
||||
index_t itmp = get_block_1d_id() - k_block_work_id * (HBlockWork * WBlockWork * NBlockWork);
|
||||
const index_t h_block_work_id = itmp / (WBlockWork * NBlockWork);
|
||||
itmp -= h_block_work_id * (WBlockWork * NBlockWork);
|
||||
const index_t w_block_work_id = itmp / NBlockWork;
|
||||
const index_t n_block_work_id = itmp - w_block_work_id * NBlockWork;
|
||||
|
||||
const index_t k_block_data_begin = k_block_work_id * KPerBlock;
|
||||
const index_t ho_block_data_begin = h_block_work_id * HoPerBlock;
|
||||
const index_t wo_block_data_begin = w_block_work_id * WoPerBlock;
|
||||
const index_t n_block_data_begin = n_block_work_id * NPerBlock;
|
||||
|
||||
const index_t hi_block_data_begin = ho_block_data_begin;
|
||||
const index_t wi_block_data_begin = wo_block_data_begin;
|
||||
|
||||
// global tensor view
|
||||
constexpr auto wei_c_k_global_desc =
|
||||
make_ConstantTensorDescriptor(Sequence<C, K>{}, Sequence<Y * X * K, 1>{});
|
||||
|
||||
// LDS tensor view
|
||||
// be careful of alignment
|
||||
constexpr index_t max_align = mod_conv::max(InBlockCopyDataPerRead_N,
|
||||
WeiBlockCopyDataPerRead_K,
|
||||
GemmDataPerReadA,
|
||||
GemmDataPerReadB);
|
||||
|
||||
constexpr auto in_c_h_w_n_block_desc = make_ConstantTensorDescriptor_aligned(
|
||||
Sequence<CPerBlock, HoPerBlock, WoPerBlock, NPerBlock>{}, Number<max_align>{});
|
||||
|
||||
constexpr auto wei_c_k_block_desc = make_ConstantTensorDescriptor_aligned(
|
||||
Sequence<CPerBlock, KPerBlock>{}, Number<max_align>{});
|
||||
|
||||
// tensor view of threadwise output in register
|
||||
constexpr auto out_k_h_w_n_thread_desc = make_ConstantTensorDescriptor(
|
||||
Sequence<KPerThread, HoPerThread, WoPerThread, NPerThread>{});
|
||||
|
||||
// blockwise copy
|
||||
// input: format is [C, Hi, Wi, N]
|
||||
#if 0
|
||||
const auto blockwise_in_copy =
|
||||
Blockwise4dTensorCopy1<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()),
|
||||
InBlockCopyDataPerRead_N>{};
|
||||
#else
|
||||
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,
|
||||
InBlockCopyDataPerRead_N>{};
|
||||
#endif
|
||||
|
||||
// blockwise wei copy
|
||||
// format is [CPerBlock, X * KPerBlock]
|
||||
const auto blockwise_wei_copy =
|
||||
Blockwise2dTensorCopy1<BlockSize,
|
||||
Float,
|
||||
decltype(wei_c_k_global_desc),
|
||||
decltype(wei_c_k_block_desc),
|
||||
decltype(wei_c_k_block_desc.GetLengths()),
|
||||
WeiBlockCopyDataPerRead_K>{};
|
||||
|
||||
// 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_c_k_block_mtx_desc = make_ConstantMatrixDescriptor(
|
||||
Number<CPerBlock>{}, Number<KPerBlock>{}, Number<wei_c_k_block_desc.GetStride(I0)>{});
|
||||
|
||||
constexpr auto b_c_wn_block_mtx_desc =
|
||||
make_ConstantMatrixDescriptor(Number<CPerBlock>{},
|
||||
Number<WoPerBlock * NPerBlock>{},
|
||||
Number<in_c_h_w_n_block_desc.GetStride(I0)>{});
|
||||
|
||||
constexpr auto c_k_wn_thread_mtx_desc =
|
||||
make_ConstantMatrixDescriptor(Number<KPerThread>{},
|
||||
Number<WoPerThread * NPerThread>{},
|
||||
Number<out_k_h_w_n_thread_desc.GetStride(I0)>{});
|
||||
|
||||
const auto blockwise_batch_gemm =
|
||||
BlockwiseBatchGemmBlockABlockBThreadCTransANormalBNormalC_V2<
|
||||
BlockSize,
|
||||
decltype(a_c_k_block_mtx_desc),
|
||||
decltype(b_c_wn_block_mtx_desc),
|
||||
decltype(c_k_wn_thread_mtx_desc),
|
||||
0,
|
||||
in_c_h_w_n_block_desc.GetStride(I1),
|
||||
out_k_h_w_n_thread_desc.GetStride(I1),
|
||||
HoPerBlock,
|
||||
GemmMPerThreadSubC,
|
||||
GemmNPerThreadSubC,
|
||||
GemmMLevel0Cluster,
|
||||
GemmNLevel0Cluster,
|
||||
GemmMLevel1Cluster,
|
||||
GemmNLevel1Cluster,
|
||||
GemmKPerThreadLoop,
|
||||
HoPerThread,
|
||||
GemmDataPerReadA,
|
||||
GemmDataPerReadB>{};
|
||||
|
||||
// LDS: be careful of alignment
|
||||
constexpr index_t in_block_space =
|
||||
in_c_h_w_n_block_desc.GetElementSpace(Number<max_align>{});
|
||||
constexpr index_t wei_block_space = wei_c_k_block_desc.GetElementSpace(Number<max_align>{});
|
||||
|
||||
__shared__ Float p_in_block[in_block_space];
|
||||
__shared__ Float p_wei_block[wei_block_space];
|
||||
|
||||
// register
|
||||
Float p_out_thread[out_k_h_w_n_thread_desc.GetElementSpace()];
|
||||
|
||||
#if 0
|
||||
if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0)
|
||||
{
|
||||
print_ConstantTensorDescriptor(in_c_h_w_n_global_desc, "in_c_h_w_n_global_desc");
|
||||
print_ConstantTensorDescriptor(wei_c_y_x_k_global_desc, "wei_c_y_x_k_global_desc");
|
||||
|
||||
print_ConstantTensorDescriptor(in_c_h_w_n_block_desc, "in_c_h_w_n_block_desc");
|
||||
print_ConstantTensorDescriptor(wei_c_x_k_block_desc, "wei_c_x_k_block_desc");
|
||||
|
||||
printf("in_block_space %u, wei_block_space %u\n", in_block_space, wei_block_space);
|
||||
}
|
||||
#endif
|
||||
|
||||
// set threadwise output tensor to 0
|
||||
threadwise_4d_tensor_set_zero(out_k_h_w_n_thread_desc, p_out_thread);
|
||||
|
||||
#if 1
|
||||
const Float* p_in_global_block_offset =
|
||||
p_in_global + in_c_h_w_n_global_desc.Get1dIndex(
|
||||
0, hi_block_data_begin, wi_block_data_begin, n_block_data_begin);
|
||||
|
||||
const Float* p_wei_global_block_offset =
|
||||
p_wei_global + wei_c_y_x_k_global_desc.Get1dIndex(0, 0, 0, k_block_data_begin);
|
||||
|
||||
for(index_t c_block_data_begin = 0; c_block_data_begin < C; c_block_data_begin += CPerBlock,
|
||||
p_in_global_block_offset += CPerBlock * in_c_h_w_n_global_desc.GetStride(I0),
|
||||
p_wei_global_block_offset += CPerBlock * wei_c_y_x_k_global_desc.GetStride(I0))
|
||||
{
|
||||
for(index_t y = 0; y < Y; ++y)
|
||||
{
|
||||
for(index_t x = 0; x < X; ++x)
|
||||
{
|
||||
blockwise_in_copy.Run(p_in_global_block_offset +
|
||||
in_c_h_w_n_global_desc.Get1dIndex(0, y, x, 0),
|
||||
p_in_block);
|
||||
|
||||
blockwise_wei_copy.Run(p_wei_global_block_offset +
|
||||
wei_c_y_x_k_global_desc.Get1dIndex(0, y, x, 0),
|
||||
p_wei_block);
|
||||
|
||||
__syncthreads();
|
||||
|
||||
blockwise_batch_gemm.Run(p_wei_block, p_in_block, p_out_thread);
|
||||
|
||||
__syncthreads();
|
||||
}
|
||||
}
|
||||
}
|
||||
#else
|
||||
for(index_t y = 0; y < Y; ++y)
|
||||
{
|
||||
for(index_t x = 0; x < X; ++x)
|
||||
{
|
||||
const Float* p_in_global_block_offset =
|
||||
p_in_global +
|
||||
in_c_h_w_n_global_desc.Get1dIndex(
|
||||
0, hi_block_data_begin + y, wi_block_data_begin + x, n_block_data_begin);
|
||||
|
||||
const Float* p_wei_global_block_offset =
|
||||
p_wei_global + wei_c_y_x_k_global_desc.Get1dIndex(0, y, x, k_block_data_begin);
|
||||
|
||||
for(index_t c_block_data_begin = 0; c_block_data_begin < C;
|
||||
c_block_data_begin += CPerBlock,
|
||||
p_in_global_block_offset +=
|
||||
CPerBlock * in_c_h_w_n_global_desc.GetStride(I0),
|
||||
p_wei_global_block_offset +=
|
||||
CPerBlock * wei_c_y_x_k_global_desc.GetStride(I0))
|
||||
{
|
||||
blockwise_in_copy.Run(p_in_global_block_offset, p_in_block);
|
||||
|
||||
blockwise_wei_copy.Run(p_wei_global_block_offset, p_wei_block);
|
||||
|
||||
__syncthreads();
|
||||
|
||||
blockwise_batch_gemm.Run(p_wei_block, p_in_block, p_out_thread);
|
||||
|
||||
__syncthreads();
|
||||
}
|
||||
}
|
||||
}
|
||||
#endif
|
||||
|
||||
// output: register to global mem,
|
||||
const auto c_thread_mtx_begin =
|
||||
blockwise_batch_gemm.GetBeginOfThreadMatrixC(get_thread_local_1d_id());
|
||||
|
||||
const index_t k_thread_data_begin = c_thread_mtx_begin.row;
|
||||
const index_t ho_thread_data_begin = c_thread_mtx_begin.batch;
|
||||
const index_t wo_thread_data_begin = c_thread_mtx_begin.col / NPerBlock;
|
||||
const index_t n_thread_data_begin =
|
||||
c_thread_mtx_begin.col - NPerBlock * wo_thread_data_begin;
|
||||
|
||||
// output is a 10d tensor
|
||||
constexpr index_t N2 = GemmNPerThreadSubC;
|
||||
constexpr index_t N1 = NPerBlock / N2;
|
||||
|
||||
constexpr index_t W2 =
|
||||
(GemmNLevel0Cluster * GemmNLevel1Cluster) / (NPerBlock / GemmNPerThreadSubC);
|
||||
constexpr index_t W1 = WoPerBlock / W2;
|
||||
|
||||
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_thread_desc = make_ConstantTensorDescriptor(
|
||||
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, 1, 1, N2>{});
|
||||
|
||||
#if 0
|
||||
if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0)
|
||||
{
|
||||
print_ConstantTensorDescriptor(out_khwn_thread_desc, "out_khwn_thread_desc");
|
||||
print_ConstantTensorDescriptor(out_10d_thread_desc, "out_10d_thread_desc");
|
||||
|
||||
print_ConstantTensorDescriptor(out_khwn_global_desc, "out_khwn_global_desc");
|
||||
print_ConstantTensorDescriptor(out_10d_global_desc, "out_10d_global_desc");
|
||||
}
|
||||
#endif
|
||||
|
||||
threadwise_10d_tensor_copy(out_10d_thread_desc,
|
||||
p_out_thread,
|
||||
out_10d_global_desc,
|
||||
p_out_global + out_k_h_w_n_global_desc.Get1dIndex(
|
||||
k_block_data_begin + k_thread_data_begin,
|
||||
ho_block_data_begin + ho_thread_data_begin,
|
||||
wo_block_data_begin + wo_thread_data_begin,
|
||||
n_block_data_begin + n_thread_data_begin),
|
||||
out_10d_thread_desc.GetLengths(),
|
||||
Number<OutThreadCopyDataPerWrite_N>{});
|
||||
}
|
||||
};
|
||||
@@ -0,0 +1,297 @@
|
||||
#pragma once
|
||||
#include "common.hip.hpp"
|
||||
#include "ConstantTensorDescriptor.hip.hpp"
|
||||
#include "ConstantMatrixDescriptor.hip.hpp"
|
||||
#include "blockwise_2d_tensor_op.hip.hpp"
|
||||
#include "blockwise_4d_tensor_op.hip.hpp"
|
||||
#include "threadwise_nd_tensor_op.hip.hpp"
|
||||
#include "threadwise_4d_tensor_op.hip.hpp"
|
||||
#include "blockwise_batched_gemm.hip.hpp"
|
||||
|
||||
template <index_t GridSize,
|
||||
index_t BlockSize,
|
||||
class Float,
|
||||
class InGlobalDesc,
|
||||
class WeiGlobalDesc,
|
||||
class OutGlobalDesc,
|
||||
index_t NPerBlock,
|
||||
index_t KPerBlock,
|
||||
index_t CPerBlock,
|
||||
index_t HoPerBlock,
|
||||
index_t WoPerBlock,
|
||||
index_t NPerThread,
|
||||
index_t KPerThread,
|
||||
index_t HoPerThread,
|
||||
index_t WoPerThread,
|
||||
index_t GemmMPerThreadSubC,
|
||||
index_t GemmNPerThreadSubC,
|
||||
index_t GemmMLevel0Cluster,
|
||||
index_t GemmNLevel0Cluster,
|
||||
index_t GemmMLevel1Cluster,
|
||||
index_t GemmNLevel1Cluster,
|
||||
index_t GemmKPerThreadLoop,
|
||||
index_t GemmDataPerReadA,
|
||||
index_t GemmDataPerReadB,
|
||||
class InBlockCopyClusterLengths_CHWN,
|
||||
index_t InBlockCopyDataPerRead_N,
|
||||
index_t WeiBlockCopyDataPerRead_K,
|
||||
index_t OutThreadCopyDataPerWrite_N>
|
||||
struct GridwiseConvolutionImplicitGemm_v1r3_lds_double_buffer_chwn_cyxk_khwn
|
||||
{
|
||||
__device__ void Run(const Float* const __restrict__ p_in_global,
|
||||
const Float* const __restrict__ p_wei_global,
|
||||
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");
|
||||
|
||||
constexpr auto I0 = Number<0>{};
|
||||
constexpr auto I1 = Number<1>{};
|
||||
constexpr auto I2 = Number<2>{};
|
||||
constexpr auto I3 = Number<3>{};
|
||||
|
||||
constexpr auto in_c_h_w_n_global_desc = InGlobalDesc{};
|
||||
constexpr auto wei_c_y_x_k_global_desc = WeiGlobalDesc{};
|
||||
constexpr auto out_k_h_w_n_global_desc = OutGlobalDesc{};
|
||||
|
||||
constexpr index_t C = in_c_h_w_n_global_desc.GetLength(I0);
|
||||
|
||||
constexpr index_t K = out_k_h_w_n_global_desc.GetLength(I0);
|
||||
constexpr index_t Ho = out_k_h_w_n_global_desc.GetLength(I1);
|
||||
constexpr index_t Wo = out_k_h_w_n_global_desc.GetLength(I2);
|
||||
constexpr index_t N = out_k_h_w_n_global_desc.GetLength(I3);
|
||||
|
||||
constexpr index_t Y = wei_c_y_x_k_global_desc.GetLength(I1);
|
||||
constexpr index_t X = wei_c_y_x_k_global_desc.GetLength(I2);
|
||||
|
||||
constexpr index_t HiPerBlock = HoPerBlock + Y - 1;
|
||||
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 ");
|
||||
|
||||
constexpr index_t KBlockWork = (K + KPerBlock - 1) / KPerBlock;
|
||||
constexpr index_t HBlockWork = (Ho + HoPerBlock - 1) / HoPerBlock;
|
||||
constexpr index_t WBlockWork = (Wo + WoPerBlock - 1) / WoPerBlock;
|
||||
constexpr index_t NBlockWork = (N + NPerBlock - 1) / NPerBlock;
|
||||
|
||||
const index_t k_block_work_id = get_block_1d_id() / (HBlockWork * WBlockWork * NBlockWork);
|
||||
index_t itmp = get_block_1d_id() - k_block_work_id * (HBlockWork * WBlockWork * NBlockWork);
|
||||
const index_t h_block_work_id = itmp / (WBlockWork * NBlockWork);
|
||||
itmp -= h_block_work_id * (WBlockWork * NBlockWork);
|
||||
const index_t w_block_work_id = itmp / NBlockWork;
|
||||
const index_t n_block_work_id = itmp - w_block_work_id * NBlockWork;
|
||||
|
||||
const index_t k_block_data_begin = k_block_work_id * KPerBlock;
|
||||
const index_t ho_block_data_begin = h_block_work_id * HoPerBlock;
|
||||
const index_t wo_block_data_begin = w_block_work_id * WoPerBlock;
|
||||
const index_t n_block_data_begin = n_block_work_id * NPerBlock;
|
||||
|
||||
const index_t hi_block_data_begin = ho_block_data_begin;
|
||||
const index_t wi_block_data_begin = wo_block_data_begin;
|
||||
|
||||
// global tensor view
|
||||
constexpr auto wei_c_k_global_desc =
|
||||
make_ConstantTensorDescriptor(Sequence<C, K>{}, Sequence<Y * X * K, 1>{});
|
||||
|
||||
// LDS tensor view
|
||||
// be careful of alignment
|
||||
constexpr index_t max_align = mod_conv::max(InBlockCopyDataPerRead_N,
|
||||
WeiBlockCopyDataPerRead_K,
|
||||
GemmDataPerReadA,
|
||||
GemmDataPerReadB);
|
||||
|
||||
constexpr auto in_c_h_w_n_block_desc = make_ConstantTensorDescriptor_aligned(
|
||||
Sequence<CPerBlock, HoPerBlock, WoPerBlock, NPerBlock>{}, Number<max_align>{});
|
||||
|
||||
constexpr auto wei_c_k_block_desc = make_ConstantTensorDescriptor_aligned(
|
||||
Sequence<CPerBlock, KPerBlock>{}, Number<max_align>{});
|
||||
|
||||
// tensor view of threadwise output in register
|
||||
constexpr auto out_k_h_w_n_thread_desc = make_ConstantTensorDescriptor(
|
||||
Sequence<KPerThread, HoPerThread, WoPerThread, NPerThread>{});
|
||||
|
||||
// blockwise copy
|
||||
// input: format is [C, Hi, Wi, N]
|
||||
#if 0
|
||||
const auto blockwise_in_copy =
|
||||
Blockwise4dTensorCopy1<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()),
|
||||
InBlockCopyDataPerRead_N>{};
|
||||
#else
|
||||
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,
|
||||
InBlockCopyDataPerRead_N>{};
|
||||
#endif
|
||||
|
||||
// blockwise wei copy
|
||||
// format is [CPerBlock, X * KPerBlock]
|
||||
const auto blockwise_wei_copy =
|
||||
Blockwise2dTensorCopy1<BlockSize,
|
||||
Float,
|
||||
decltype(wei_c_k_global_desc),
|
||||
decltype(wei_c_k_block_desc),
|
||||
decltype(wei_c_k_block_desc.GetLengths()),
|
||||
WeiBlockCopyDataPerRead_K>{};
|
||||
|
||||
// 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_c_k_block_mtx_desc = make_ConstantMatrixDescriptor(
|
||||
Number<CPerBlock>{}, Number<KPerBlock>{}, Number<wei_c_k_block_desc.GetStride(I0)>{});
|
||||
|
||||
constexpr auto b_c_wn_block_mtx_desc =
|
||||
make_ConstantMatrixDescriptor(Number<CPerBlock>{},
|
||||
Number<WoPerBlock * NPerBlock>{},
|
||||
Number<in_c_h_w_n_block_desc.GetStride(I0)>{});
|
||||
|
||||
constexpr auto c_k_wn_thread_mtx_desc =
|
||||
make_ConstantMatrixDescriptor(Number<KPerThread>{},
|
||||
Number<WoPerThread * NPerThread>{},
|
||||
Number<out_k_h_w_n_thread_desc.GetStride(I0)>{});
|
||||
|
||||
const auto blockwise_batch_gemm =
|
||||
BlockwiseBatchGemmBlockABlockBThreadCTransANormalBNormalC_V2<
|
||||
BlockSize,
|
||||
decltype(a_c_k_block_mtx_desc),
|
||||
decltype(b_c_wn_block_mtx_desc),
|
||||
decltype(c_k_wn_thread_mtx_desc),
|
||||
0,
|
||||
in_c_h_w_n_block_desc.GetStride(I1),
|
||||
out_k_h_w_n_thread_desc.GetStride(I1),
|
||||
HoPerBlock,
|
||||
GemmMPerThreadSubC,
|
||||
GemmNPerThreadSubC,
|
||||
GemmMLevel0Cluster,
|
||||
GemmNLevel0Cluster,
|
||||
GemmMLevel1Cluster,
|
||||
GemmNLevel1Cluster,
|
||||
GemmKPerThreadLoop,
|
||||
HoPerThread,
|
||||
GemmDataPerReadA,
|
||||
GemmDataPerReadB>{};
|
||||
|
||||
// LDS: be careful of alignment
|
||||
constexpr index_t in_block_space =
|
||||
in_c_h_w_n_block_desc.GetElementSpace(Number<max_align>{});
|
||||
constexpr index_t wei_block_space = wei_c_k_block_desc.GetElementSpace(Number<max_align>{});
|
||||
|
||||
__shared__ Float p_in_block[in_block_space];
|
||||
__shared__ Float p_wei_block[wei_block_space];
|
||||
|
||||
// register
|
||||
Float p_out_thread[out_k_h_w_n_thread_desc.GetElementSpace()];
|
||||
|
||||
#if 0
|
||||
if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0)
|
||||
{
|
||||
print_ConstantTensorDescriptor(in_c_h_w_n_global_desc, "in_c_h_w_n_global_desc");
|
||||
print_ConstantTensorDescriptor(wei_c_y_x_k_global_desc, "wei_c_y_x_k_global_desc");
|
||||
|
||||
print_ConstantTensorDescriptor(in_c_h_w_n_block_desc, "in_c_h_w_n_block_desc");
|
||||
print_ConstantTensorDescriptor(wei_c_x_k_block_desc, "wei_c_x_k_block_desc");
|
||||
|
||||
printf("in_block_space %u, wei_block_space %u\n", in_block_space, wei_block_space);
|
||||
}
|
||||
#endif
|
||||
|
||||
// set threadwise output tensor to 0
|
||||
threadwise_4d_tensor_set_zero(out_k_h_w_n_thread_desc, p_out_thread);
|
||||
|
||||
for(index_t y = 0; y < Y; ++y)
|
||||
{
|
||||
for(index_t x = 0; x < X; ++x)
|
||||
{
|
||||
const Float* p_in_global_block_offset =
|
||||
p_in_global +
|
||||
in_c_h_w_n_global_desc.Get1dIndex(
|
||||
0, hi_block_data_begin + y, wi_block_data_begin + x, n_block_data_begin);
|
||||
|
||||
const Float* p_wei_global_block_offset =
|
||||
p_wei_global + wei_c_y_x_k_global_desc.Get1dIndex(0, y, x, k_block_data_begin);
|
||||
|
||||
for(index_t c_block_data_begin = 0; c_block_data_begin < C;
|
||||
c_block_data_begin += CPerBlock,
|
||||
p_in_global_block_offset +=
|
||||
CPerBlock * in_c_h_w_n_global_desc.GetStride(I0),
|
||||
p_wei_global_block_offset +=
|
||||
CPerBlock * wei_c_y_x_k_global_desc.GetStride(I0))
|
||||
{
|
||||
blockwise_in_copy.Run(p_in_global_block_offset, p_in_block);
|
||||
|
||||
blockwise_wei_copy.Run(p_wei_global_block_offset, p_wei_block);
|
||||
|
||||
__syncthreads();
|
||||
|
||||
blockwise_batch_gemm.Run(p_wei_block, p_in_block, p_out_thread);
|
||||
|
||||
__syncthreads();
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// output: register to global mem,
|
||||
const auto c_thread_mtx_begin =
|
||||
blockwise_batch_gemm.GetBeginOfThreadMatrixC(get_thread_local_1d_id());
|
||||
|
||||
const index_t k_thread_data_begin = c_thread_mtx_begin.row;
|
||||
const index_t ho_thread_data_begin = c_thread_mtx_begin.batch;
|
||||
const index_t wo_thread_data_begin = c_thread_mtx_begin.col / NPerBlock;
|
||||
const index_t n_thread_data_begin =
|
||||
c_thread_mtx_begin.col - NPerBlock * wo_thread_data_begin;
|
||||
|
||||
// output is a 10d tensor
|
||||
constexpr index_t N2 = GemmNPerThreadSubC;
|
||||
constexpr index_t N1 = NPerBlock / N2;
|
||||
|
||||
constexpr index_t W2 =
|
||||
(GemmNLevel0Cluster * GemmNLevel1Cluster) / (NPerBlock / GemmNPerThreadSubC);
|
||||
constexpr index_t W1 = WoPerBlock / W2;
|
||||
|
||||
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_thread_desc = make_ConstantTensorDescriptor(
|
||||
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, 1, 1, N2>{});
|
||||
|
||||
#if 0
|
||||
if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0)
|
||||
{
|
||||
print_ConstantTensorDescriptor(out_khwn_thread_desc, "out_khwn_thread_desc");
|
||||
print_ConstantTensorDescriptor(out_10d_thread_desc, "out_10d_thread_desc");
|
||||
|
||||
print_ConstantTensorDescriptor(out_khwn_global_desc, "out_khwn_global_desc");
|
||||
print_ConstantTensorDescriptor(out_10d_global_desc, "out_10d_global_desc");
|
||||
}
|
||||
#endif
|
||||
|
||||
threadwise_10d_tensor_copy(out_10d_thread_desc,
|
||||
p_out_thread,
|
||||
out_10d_global_desc,
|
||||
p_out_global + out_k_h_w_n_global_desc.Get1dIndex(
|
||||
k_block_data_begin + k_thread_data_begin,
|
||||
ho_block_data_begin + ho_thread_data_begin,
|
||||
wo_block_data_begin + wo_thread_data_begin,
|
||||
n_block_data_begin + n_thread_data_begin),
|
||||
out_10d_thread_desc.GetLengths(),
|
||||
Number<OutThreadCopyDataPerWrite_N>{});
|
||||
}
|
||||
};
|
||||
Reference in New Issue
Block a user