Skip to content

GitLab

Commit 8fcf3f1e authored by Chao Liu's avatar Chao Liu
Browse files

added implicit GEMM v3

parent 8ce14804
Hide whitespace changes
Inline Side-by-side
Showing with 100 additions and 106 deletions
driver/device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw.hpp
View file @ 8fcf3f1e
...@@ -57,7 +57,7 @@ void device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw(InDesc, ...@@ -57,7 +57,7 @@ void device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw(InDesc,
wei_cyxk_device_buf.ToDevice(wei_cyxk.mData.data()); wei_cyxk_device_buf.ToDevice(wei_cyxk.mData.data());
out_nkhw_device_buf.ToDevice(out_nkhw.mData.data()); out_nkhw_device_buf.ToDevice(out_nkhw.mData.data());
#if 1 #if 0
// for 3x3, 34x34, v1r3, Pascal // for 3x3, 34x34, v1r3, Pascal
constexpr index_t BlockSize = 128; constexpr index_t BlockSize = 128;
...@@ -127,7 +127,7 @@ void device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw(InDesc, ...@@ -127,7 +127,7 @@ void device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw(InDesc,
constexpr index_t WeiBlockCopyDataPerRead_K = 4; constexpr index_t WeiBlockCopyDataPerRead_K = 4;
constexpr index_t OutThreadCopyDataPerWrite_W = 4; constexpr index_t OutThreadCopyDataPerWrite_W = 4;
#elif 0 #elif 1
// for 3x3, 34x34, v1r3, Vega 20, WoPerBlock = 16 // for 3x3, 34x34, v1r3, Vega 20, WoPerBlock = 16
constexpr index_t BlockSize = 256; constexpr index_t BlockSize = 256;
...@@ -313,7 +313,7 @@ void device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw(InDesc, ...@@ -313,7 +313,7 @@ void device_convolution_implicit_gemm_v1_nchw_cyxk_nkhw(InDesc,
for(index_t i = 0; i < nrepeat; ++i) for(index_t i = 0; i < nrepeat; ++i)
{ {
constexpr auto gridwise_conv = constexpr auto gridwise_conv =
#if 0 #if 1
GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
#else #else
GridwiseConvolutionImplicitGemm_v1r3_lds_double_buffer_nchw_cyxk_nkhw GridwiseConvolutionImplicitGemm_v1r3_lds_double_buffer_nchw_cyxk_nkhw
......
driver/device_convolution_implicit_gemm_v3_nchw_cyxk_nkhw.hpp
View file @ 8fcf3f1e
...@@ -62,7 +62,6 @@ void device_convolution_implicit_gemm_v3_nchw_cyxk_nkhw(InDesc, ...@@ -62,7 +62,6 @@ void device_convolution_implicit_gemm_v3_nchw_cyxk_nkhw(InDesc,
constexpr index_t B = (N * Ho * Wo) / (N1 * N2); constexpr index_t B = (N * Ho * Wo) / (N1 * N2);
#if 1 #if 1
// for 3x3, 28x28, v3
constexpr index_t BlockSize = 256; constexpr index_t BlockSize = 256;
constexpr index_t BPerBlock = 16; constexpr index_t BPerBlock = 16;
......
src/include/gridwise_convolution_implicit_gemm_v1r3_nchw_cyxk_nkhw.hip.hpp
View file @ 8fcf3f1e
...@@ -83,7 +83,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -83,7 +83,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
constexpr index_t HBlockWork = mod_conv::integer_divide_ceil(Ho, HoPerBlock); constexpr index_t HBlockWork = mod_conv::integer_divide_ceil(Ho, HoPerBlock);
constexpr index_t WBlockWork = mod_conv::integer_divide_ceil(Wo, WoPerBlock); constexpr index_t WBlockWork = mod_conv::integer_divide_ceil(Wo, WoPerBlock);
constexpr auto block_work_desc = make_ConstantTensorDescriptor( constexpr auto block_work_desc = make_ConstantTensorDescriptor_default_rank_packed(
Sequence<NBlockWork, KBlockWork, HBlockWork, WBlockWork>{}); Sequence<NBlockWork, KBlockWork, HBlockWork, WBlockWork>{});
const auto block_work_multi_id = const auto block_work_multi_id =
...@@ -99,7 +99,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -99,7 +99,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
// global tensor view // global tensor view
constexpr auto wei_c_k_global_desc = constexpr auto wei_c_k_global_desc =
make_ConstantTensorDescriptor(Sequence<C, K>{}, Sequence<Y * X * K, 1>{}); make_ConstantTensorDescriptor_default_rank(Sequence<C, K>{}, Sequence<Y * X * K, 1>{});
// LDS tensor view // LDS tensor view
// be careful of alignment // be careful of alignment
...@@ -108,7 +108,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -108,7 +108,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
GemmDataPerReadA, GemmDataPerReadA,
GemmDataPerReadB); GemmDataPerReadB);
constexpr auto in_c_h_w_n_block_desc = make_ConstantTensorDescriptor_aligned( constexpr auto in_c_h_w_n_block_desc = make_ConstantTensorDescriptor_default_rank_aligned(
Sequence<CPerBlock, HoPerBlock, WoPerBlock, NPerBlock>{}, Sequence<CPerBlock, HoPerBlock, WoPerBlock, NPerBlock>{},
Number<InBlockReorderDataPerWrite_N>{}); Number<InBlockReorderDataPerWrite_N>{});
...@@ -117,12 +117,12 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -117,12 +117,12 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
static_assert(in_c_h_w_n_block_desc.GetStride(I1) % GemmDataPerReadB == 0, static_assert(in_c_h_w_n_block_desc.GetStride(I1) % GemmDataPerReadB == 0,
"GemmDataPerReadB alignment requirement is not meet"); "GemmDataPerReadB alignment requirement is not meet");
constexpr auto wei_c_k_block_desc = make_ConstantTensorDescriptor_aligned( constexpr auto wei_c_k_block_desc = make_ConstantTensorDescriptor_default_rank_aligned(
Sequence<CPerBlock, KPerBlock>{}, Sequence<CPerBlock, KPerBlock>{},
Number<mod_conv::max(WeiBlockCopyDataPerRead_K, GemmDataPerReadA)>{}); Number<mod_conv::max(WeiBlockCopyDataPerRead_K, GemmDataPerReadA)>{});
// tensor view of threadwise output in register // tensor view of threadwise output in register
constexpr auto out_k_h_w_n_thread_desc = make_ConstantTensorDescriptor( constexpr auto out_k_h_w_n_thread_desc = make_ConstantTensorDescriptor_default_rank_packed(
Sequence<KPerThread, HoPerThread, WoPerThread, NPerThread>{}); Sequence<KPerThread, HoPerThread, WoPerThread, NPerThread>{});
// blockwise copy // blockwise copy
...@@ -140,7 +140,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -140,7 +140,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
decltype(map_chwn2nchw), decltype(map_chwn2nchw),
InBlockReorderMapThreadCluster2SrcCluster_CHNW2NCHW, InBlockReorderMapThreadCluster2SrcCluster_CHNW2NCHW,
InBlockReorderDataPerRead_W, InBlockReorderDataPerRead_W,
InBlockReorderDataPerWrite_N>{}; InBlockReorderDataPerWrite_N>({0, 0, 0, 0}, {0, 0, 0, 0});
// blockwise wei copy // blockwise wei copy
// format is [CPerBlock, KPerBlock] // format is [CPerBlock, KPerBlock]
...@@ -150,7 +150,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -150,7 +150,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
decltype(wei_c_k_global_desc), decltype(wei_c_k_global_desc),
decltype(wei_c_k_block_desc), decltype(wei_c_k_block_desc),
decltype(wei_c_k_block_desc.GetLengths()), decltype(wei_c_k_block_desc.GetLengths()),
WeiBlockCopyDataPerRead_K>{}; WeiBlockCopyDataPerRead_K>({0, 0}, {0, 0});
// a series of blockwise batched GEMM // a series of blockwise batched GEMM
// C_matrix += transpose(A_matrix) * B_matrix // C_matrix += transpose(A_matrix) * B_matrix
...@@ -194,7 +194,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -194,7 +194,7 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
// choose GEMM implementation here // choose GEMM implementation here
const auto run_blockwise_batch_gemm = [&](auto... Xs) { const auto run_blockwise_batch_gemm = [&](auto... Xs) {
#if 0 #if 1
return blockwise_batch_gemm.Run(Xs...); return blockwise_batch_gemm.Run(Xs...);
#elif 0 #elif 0
return blockwise_batch_gemm.Run_asm(Xs...); return blockwise_batch_gemm.Run_asm(Xs...);
...@@ -249,7 +249,6 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -249,7 +249,6 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
{ {
for(index_t x = 0; x < X; ++x) for(index_t x = 0; x < X; ++x)
{ {
#if 1
blockwise_in_copy_reorder.Run(p_in_global_block_offset + blockwise_in_copy_reorder.Run(p_in_global_block_offset +
in_n_c_h_w_global_desc.GetOffsetFromMultiIndex(0, 0, y, x), in_n_c_h_w_global_desc.GetOffsetFromMultiIndex(0, 0, y, x),
p_in_block); p_in_block);
...@@ -257,23 +256,6 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -257,23 +256,6 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
blockwise_wei_copy.Run(p_wei_global_block_offset + blockwise_wei_copy.Run(p_wei_global_block_offset +
wei_c_y_x_k_global_desc.GetOffsetFromMultiIndex(0, y, x, 0), wei_c_y_x_k_global_desc.GetOffsetFromMultiIndex(0, y, x, 0),
p_wei_block); p_wei_block);
#else
Float p_in_clipboard[blockwise_in_copy_reorder.GetRegisterClipboardSize()];
Float p_wei_clipboard[blockwise_wei_copy.GetRegisterClipboardSize()];
blockwise_in_copy_reorder.RunLoadRegisterClipboard(
p_in_global_block_offset + in_n_c_h_w_global_desc.GetOffsetFromMultiIndex(0, 0, y, x),
p_in_clipboard);
blockwise_wei_copy.RunLoadRegisterClipboard(
p_wei_global_block_offset + wei_c_y_x_k_global_desc.GetOffsetFromMultiIndex(0, y, x, 0),
p_wei_clipboard);
blockwise_wei_copy.RunStoreRegisterClipboard(p_wei_clipboard, p_wei_block);
blockwise_in_copy_reorder.RunStoreRegisterClipboard(p_in_clipboard, p_in_block);
#endif
__syncthreads(); __syncthreads();
...@@ -304,24 +286,9 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -304,24 +286,9 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
p_wei_global_block_offset += p_wei_global_block_offset +=
CPerBlock * wei_c_y_x_k_global_desc.GetStride(I0)) CPerBlock * wei_c_y_x_k_global_desc.GetStride(I0))
{ {
#if 0 blockwise_in_copy_reorder.Run(p_in_global_block_offset, p_in_block);
blockwise_in_copy_reorder.Run(p_in_global_block_offset,
p_in_block);
blockwise_wei_copy.Run(p_wei_global_block_offset,
p_wei_block);
#else
Float p_in_clipboard[blockwise_in_copy_reorder.GetRegisterClipboardSize()];
Float p_wei_clipboard[blockwise_wei_copy.GetRegisterClipboardSize()];
blockwise_in_copy_reorder.RunLoadRegisterClipboard(p_in_global_block_offset,
p_in_clipboard);
blockwise_wei_copy.RunLoadRegisterClipboard(p_wei_global_block_offset,
p_wei_clipboard);
blockwise_wei_copy.RunStoreRegisterClipboard(p_wei_clipboard, p_wei_block); blockwise_wei_copy.Run(p_wei_global_block_offset, p_wei_block);
blockwise_in_copy_reorder.RunStoreRegisterClipboard(p_in_clipboard, p_in_block);
#endif
__syncthreads(); __syncthreads();
...@@ -342,13 +309,12 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -342,13 +309,12 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
const index_t wo_thread_data_begin = c_thread_mtx_begin.col / NPerBlock; 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; const index_t n_thread_data_begin = c_thread_mtx_begin.col % NPerBlock;
static_if<GemmNPerThreadSubC <= NPerBlock>{}([&](auto f_dummy) { // f_dummy do nothing but static_if<GemmNPerThreadSubC <= NPerBlock>{}([&](auto fwd) {
// perfect forwarding. // fwd do nothing but perfect forwarding.
// Using this trick to // Using this trick to make this lambda a generic lambda, so it won't be compiled until
// make this lambda a generic lambda, so it won't be compiled until // begin instantiated here
// instantiated
static_assert( static_assert(
(f_dummy(GemmNPerThreadSubC) <= NPerBlock && NPerBlock % GemmNPerThreadSubC == 0), (fwd(GemmNPerThreadSubC) <= NPerBlock && NPerBlock % GemmNPerThreadSubC == 0),
"wrong!"); "wrong!");
// output is a 10d tensor // output is a 10d tensor
...@@ -356,38 +322,33 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -356,38 +322,33 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
constexpr index_t N1 = NPerBlock / N2; constexpr index_t N1 = NPerBlock / N2;
constexpr index_t W2 = constexpr index_t W2 =
(GemmNLevel0Cluster * GemmNLevel1Cluster) / f_dummy(NPerBlock / GemmNPerThreadSubC); (GemmNLevel0Cluster * GemmNLevel1Cluster) / fwd(NPerBlock / GemmNPerThreadSubC);
constexpr index_t W1 = WoPerBlock / W2; constexpr index_t W1 = WoPerBlock / W2;
constexpr index_t K2 = GemmMPerThreadSubC; constexpr index_t K2 = GemmMPerThreadSubC;
constexpr index_t K1 = KPerBlock / KPerThread; constexpr index_t K1 = KPerBlock / KPerThread;
constexpr auto out_10d_global_desc = constexpr auto out_10d_global_desc = fwd(out_n_k_h_w_global_desc)
make_ConstantTensorDescriptor(Sequence<N / f_dummy(N1 * N2), .Fold(I3, Number<W1>{}, Number<W2>{})
N1, .Fold(I1, Number<K1>{}, Number<K2>{})
N2, .Fold(I0, Number<N1>{}, Number<N2>{});
K / (K1 * K2),
K1, constexpr auto out_10d_thread_desc = fwd(out_k_h_w_n_thread_desc)
K2, .Fold(I3, Number<1>{}, Number<N2>{})
Ho, .Fold(I2, Number<W1>{}, Number<1>{})
Wo / (W1 * W2), .Fold(I0, Number<1>{}, Number<K2>{});
W1,
W2>{});
constexpr auto out_10d_thread_desc = make_ConstantTensorDescriptor(
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, 1, 1, N2>{});
#if 0 #if 0
if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0) if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0)
{ {
print_ConstantTensorDescriptor(out_k_h_w_n_thread_desc, print_ConstantTensorDescriptor(out_k_h_w_n_thread_desc,
"out_k_h_w_n_thread_desc"); "a: out_k_h_w_n_thread_desc");
print_ConstantTensorDescriptor(out_10d_thread_desc, "out_10d_thread_desc"); print_ConstantTensorDescriptor(out_10d_thread_desc, "a: out_10d_thread_desc");
print_ConstantTensorDescriptor(out_k_h_w_n_global_desc, print_ConstantTensorDescriptor(out_n_k_h_w_global_desc,
"out_k_h_w_n_global_desc"); "a: out_n_k_h_w_global_desc");
print_ConstantTensorDescriptor(out_10d_global_desc, "out_10d_global_desc"); print_ConstantTensorDescriptor(out_10d_global_desc, "a: out_10d_global_desc");
} }
#endif #endif
constexpr auto map_out_global2thread = Sequence<7, 8, 9, 0, 1, 2, 3, 4, 5, 6>{}; constexpr auto map_out_global2thread = Sequence<7, 8, 9, 0, 1, 2, 3, 4, 5, 6>{};
...@@ -405,8 +366,8 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -405,8 +366,8 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
out_10d_thread_desc.GetLengths(), out_10d_thread_desc.GetLengths(),
map_out_global2thread); map_out_global2thread);
// Number<OutThreadCopyDataPerWrite_W>{}); // Number<OutThreadCopyDataPerWrite_W>{});
}).else_([&](auto f_dummy) { }).else_([&](auto fwd) {
static_assert(f_dummy(GemmNPerThreadSubC) >= NPerBlock && NPerThread == NPerBlock && static_assert(fwd(GemmNPerThreadSubC) >= NPerBlock && NPerThread == NPerBlock &&
GemmNPerThreadSubC % NPerThread == 0, GemmNPerThreadSubC % NPerThread == 0,
"wrong!"); "wrong!");
...@@ -415,34 +376,40 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -415,34 +376,40 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
constexpr index_t W3 = GemmNPerThreadSubC / NPerBlock; constexpr index_t W3 = GemmNPerThreadSubC / NPerBlock;
constexpr index_t W2 = GemmNLevel0Cluster * GemmNLevel1Cluster; constexpr index_t W2 = GemmNLevel0Cluster * GemmNLevel1Cluster;
constexpr index_t W1 = WoPerBlock / f_dummy(W2 * W3); constexpr index_t W1 = WoPerBlock / fwd(W2 * W3);
constexpr index_t K2 = GemmMPerThreadSubC; constexpr index_t K2 = GemmMPerThreadSubC;
constexpr index_t K1 = KPerBlock / KPerThread; constexpr index_t K1 = KPerBlock / KPerThread;
constexpr auto out_10d_global_desc = make_ConstantTensorDescriptor( constexpr auto out_10d_global_desc =
Sequence<N / N1, N1, K / (K1 * K2), K1, K2, Ho, Wo / (W1 * W2 * W3), W1, W2, W3>{}); fwd(out_n_k_h_w_global_desc)
.Fold(I3, Number<W1>{}, Number<W2>{}, Number<W3>{})
.Fold(I1, Number<K1>{}, Number<K2>{})
.Fold(I0, Number<N1>{});
constexpr auto out_10d_thread_desc = make_ConstantTensorDescriptor( constexpr auto out_10d_thread_desc =
Sequence<KPerThread / K2, 1, K2, HoPerThread, 1, W1, 1, W3, 1, N1>{}); fwd(out_k_h_w_n_thread_desc)
.Fold(I3, Number<N1>{})
.Fold(I2, Number<W1>{}, Number<1>{}, Number<W3>{})
.Fold(I0, Number<1>{}, Number<K2>{});
#if 0 #if 0
if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0) if(get_thread_local_1d_id() == 0 && get_block_1d_id() == 0)
{ {
print_ConstantTensorDescriptor(out_k_h_w_n_thread_desc, print_ConstantTensorDescriptor(out_k_h_w_n_thread_desc,
"out_k_h_w_n_thread_desc"); "b: out_k_h_w_n_thread_desc");
print_ConstantTensorDescriptor(out_10d_thread_desc, "out_10d_thread_desc"); print_ConstantTensorDescriptor(out_10d_thread_desc, "b: out_10d_thread_desc");
print_ConstantTensorDescriptor(out_k_h_w_n_global_desc,
"out_k_h_w_n_global_desc");
print_ConstantTensorDescriptor(out_10d_global_desc, "out_10d_global_desc");
} print_ConstantTensorDescriptor(out_n_k_h_w_global_desc,
"b: out_n_k_h_w_global_desc");
print_ConstantTensorDescriptor(out_10d_global_desc, "b: out_10d_global_desc");
}
#endif #endif
constexpr auto map_out_global2thread = Sequence<8, 9, 0, 1, 2, 3, 4, 5, 6, 7>{}; constexpr auto map_out_global2thread = Sequence<8, 9, 0, 1, 2, 3, 4, 5, 6, 7>{};
threadwise_tensor_slice_copy_reorder_given_dst2src_v2( #if 0
threadwise_tensor_slice_copy_reorder_given_dst2src_v3(
out_10d_thread_desc, out_10d_thread_desc,
p_out_thread, p_out_thread,
out_10d_global_desc, out_10d_global_desc,
...@@ -453,8 +420,24 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw ...@@ -453,8 +420,24 @@ struct GridwiseConvolutionImplicitGemm_v1r3_nchw_cyxk_nkhw
ho_block_data_begin + ho_thread_data_begin, ho_block_data_begin + ho_thread_data_begin,
wo_block_data_begin + wo_thread_data_begin), wo_block_data_begin + wo_thread_data_begin),
out_10d_thread_desc.GetLengths(), out_10d_thread_desc.GetLengths(),
map_out_global2thread); map_out_global2thread,
// Number<OutThreadCopyDataPerWrite_W>{}); Number<OutThreadCopyDataPerWrite_W>{});
#else
threadwise_tensor_slice_copy_generic(
out_10d_thread_desc.ReorderGivenNew2Old(map_out_global2thread),
p_out_thread,
make_zero_array<index_t, 10>(),
out_10d_global_desc,
p_out_global +
out_n_k_h_w_global_desc.GetOffsetFromMultiIndex(
n_block_data_begin + n_thread_data_begin,
k_block_data_begin + k_thread_data_begin,
ho_block_data_begin + ho_thread_data_begin,
wo_block_data_begin + wo_thread_data_begin),
make_zero_array<index_t, 10>(),
out_10d_thread_desc.GetLengths().ReorderGivenNew2Old(map_out_global2thread),
arithmetic_sequence_gen<0, 10, 1>::SeqType{});
#endif
}); });
} }
}; };
src/include/gridwise_convolution_implicit_gemm_v3_nchw_cyxk_nkhw.hip.hpp
View file @ 8fcf3f1e
...@@ -151,6 +151,7 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw ...@@ -151,6 +151,7 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw
// slice a tensor, and copy it into another tensor // slice a tensor, and copy it into another tensor
// this copy operator already have blockwise offset built-in // this copy operator already have blockwise offset built-in
const auto blockwise_wei_copy = const auto blockwise_wei_copy =
#if 0
BlockwiseTensorSliceCopy_generic_v1<BlockSize, BlockwiseTensorSliceCopy_generic_v1<BlockSize,
Float, Float,
decltype(wei_c_k_global_desc), decltype(wei_c_k_global_desc),
...@@ -164,15 +165,26 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw ...@@ -164,15 +165,26 @@ struct GridwiseConvolutionImplicitGemm_v3_nchw_cyxk_nkhw
WeiBlockCopyDataPerAccess_K, WeiBlockCopyDataPerAccess_K,
WeiBlockCopyDataPerAccess_K>( WeiBlockCopyDataPerAccess_K>(
{0, k_block_data_on_global}, {0, 0}); {0, k_block_data_on_global}, {0, 0});
#else
// GEMM definition Blockwise2dTensorCopy3<BlockSize,
// c_mtx += transpose(a_mtx) * b_mtx Float,
// a_mtx[CPerBlock, KPerBlock] is in LDS decltype(wei_c_k_global_desc),
// b_mtx[CPerBlocl, N1 * BPerBlock * N2] is in LDS decltype(wei_c_k_block_desc),
// c_mtx[KPerBlock, N1 * BPerBlock * N2] is distributed among threads, and saved in decltype(wei_c_k_block_desc.GetLengths()),
// register WeiBlockCopyDataPerAccess_K>({0, k_block_data_on_global},
constexpr auto a_c_k_block_mtx_desc = make_ConstantMatrixDescriptor( {0, 0});
Number<CPerBlock>{}, Number<KPerBlock>{}, Number<wei_c_k_block_desc.GetStride(I0)>{}); #endif
// GEMM definition
// c_mtx += transpose(a_mtx) * b_mtx
// a_mtx[CPerBlock, KPerBlock] is in LDS
// b_mtx[CPerBlocl, N1 * BPerBlock * N2] is in LDS
// c_mtx[KPerBlock, N1 * BPerBlock * N2] is distributed among threads, and saved in
// register
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_n1bn2_block_mtx_desc = constexpr auto b_c_n1bn2_block_mtx_desc =
make_ConstantMatrixDescriptor(Number<CPerBlock>{}, make_ConstantMatrixDescriptor(Number<CPerBlock>{},
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment