Skip to content

GitLab

"git@developer.sourcefind.cn:modelzoo/resnet50_tensorflow.git" did not exist on "f2c76e41f668b30cc6714ed8a866c07cc2f2275a"
Commit 4eb9a7a4 authored by Jing Zhang's avatar Jing Zhang
Browse files

enable static desc

parent a69937d3
Hide whitespace changes
Inline Side-by-side
Showing with 142 additions and 63 deletions
composable_kernel/include/tensor_operation/gridwise_gemm_dlops_v2_add.hpp
View file @ 4eb9a7a4
...@@ -119,6 +119,52 @@ __global__ void ...@@ -119,6 +119,52 @@ __global__ void
c_blockid_to_k_n_h_w_block_cluster_adaptor, c_blockid_to_k_n_h_w_block_cluster_adaptor,
integral_constant<bool, HasMainE0BlockLoop>{}); integral_constant<bool, HasMainE0BlockLoop>{});
} }
#elif CK_EXPERIMENTAL_STATIC_TENSOR_DESCRIPTOR
template <typename GridwiseGemm,
typename FloatAB,
typename FloatC,
typename AGridDesc_E0_E1_K0_K1_E2,
typename BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2,
typename CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2,
typename DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx,
typename CBlockIdToBlockClusterAdaptor_K_N_H_W,
bool HasMainE0BlockLoop>
__global__ void
#if CK_USE_LAUNCH_BOUNDS
__launch_bounds__(CK_MAX_THREAD_PER_BLOCK, CK_MIN_BLOCK_PER_CU)
#endif
kernel_gemm_dlops_v2_add(const FloatAB* __restrict__ p_a_grid,
const FloatAB* __restrict__ p_b_grid,
const FloatC* __restrict__ p_bias_grid,
FloatC* __restrict__ p_c_grid,
FloatC* __restrict__ p_d_grid)
{
constexpr index_t shared_block_size =
GridwiseGemm::GetSharedMemoryNumberOfByte() / sizeof(FloatAB);
__shared__ FloatAB p_shared_block[shared_block_size];
constexpr auto a_e0_e1_k0_k1_e2_grid_desc = AGridDesc_E0_E1_K0_K1_E2{};
constexpr auto b_e0_e1_n_h0_h1_h2_w0_w1_w2_e2_grid_desc =
BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2{};
constexpr auto c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc = CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2{};
constexpr auto d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc = DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx{};
constexpr auto c_blockid_to_k_n_h_w_block_cluster_adaptor =
CBlockIdToBlockClusterAdaptor_K_N_H_W{};
GridwiseGemm::Run(p_a_grid,
p_b_grid,
p_bias_grid,
p_c_grid,
p_d_grid,
p_shared_block,
a_e0_e1_k0_k1_e2_grid_desc,
b_e0_e1_n_h0_h1_h2_w0_w1_w2_e2_grid_desc,
c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc,
d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc,
c_blockid_to_k_n_h_w_block_cluster_adaptor,
integral_constant<bool, HasMainE0BlockLoop>{});
}
#endif #endif
template <index_t BlockSize, template <index_t BlockSize,
...@@ -339,13 +385,23 @@ struct GridwiseGemmDlops_km_kn_mn_v3_add ...@@ -339,13 +385,23 @@ struct GridwiseGemmDlops_km_kn_mn_v3_add
const auto K1 = Number<KPerBlock>{}; const auto K1 = Number<KPerBlock>{};
const auto K0 = K / K1; const auto K0 = K / K1;
const auto H2 = HoPerThread / 2; #if CK_EXPERIMENTAL_STATIC_TENSOR_DESCRIPTOR
const auto H2 = Number<HoPerThread / 2>{};
const auto H1 = Number<HoPerBlock / HoPerThread>{}; const auto H1 = Number<HoPerBlock / HoPerThread>{};
const auto H0 = Number<Hx / (H1 * H2)>{};
const auto W2 = Number<WoPerThread / 2>{};
const auto W1 = Number<WoPerBlock / WoPerThread>{};
const auto W0 = Number<Wx / (W1 * W2)>{};
#else
const auto H2 = HoPerThread / 2;
const auto H1 = HoPerBlock / HoPerThread;
const auto H0 = Hx / (H1 * H2); const auto H0 = Hx / (H1 * H2);
const auto W2 = WoPerThread / 2; const auto W2 = WoPerThread / 2;
const auto W1 = Number<WoPerBlock / WoPerThread>{}; const auto W1 = WoPerBlock / WoPerThread;
const auto W0 = Wx / (W1 * W2); const auto W0 = Wx / (W1 * W2);
#endif
const auto d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc = transform_tensor_descriptor( const auto d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc = transform_tensor_descriptor(
d_k_n_hx_wx_grid_desc, d_k_n_hx_wx_grid_desc,
...@@ -415,10 +471,17 @@ struct GridwiseGemmDlops_km_kn_mn_v3_add ...@@ -415,10 +471,17 @@ struct GridwiseGemmDlops_km_kn_mn_v3_add
const auto Ho = c_k_n_ho_wo_grid_desc.GetLength(I2); const auto Ho = c_k_n_ho_wo_grid_desc.GetLength(I2);
const auto Wo = c_k_n_ho_wo_grid_desc.GetLength(I3); const auto Wo = c_k_n_ho_wo_grid_desc.GetLength(I3);
#if CK_EXPERIMENTAL_STATIC_TENSOR_DESCRIPTOR
const auto K0 = Number<K / KPerBlock>{};
const auto N0 = Number<N / NPerBlock>{};
const auto H0 = Number<Ho / HoPerBlock>{};
const auto W0 = Number<Wo / WoPerBlock>{};
#else
const auto K0 = K / KPerBlock; const auto K0 = K / KPerBlock;
const auto N0 = N / NPerBlock; const auto N0 = N / NPerBlock;
const auto H0 = Ho / HoPerBlock; const auto H0 = Ho / HoPerBlock;
const auto W0 = Wo / WoPerBlock; const auto W0 = Wo / WoPerBlock;
#endif
const auto c_blockid_to_k_n_ho_wo_block_cluster_adaptor = make_single_stage_tensor_adaptor( const auto c_blockid_to_k_n_ho_wo_block_cluster_adaptor = make_single_stage_tensor_adaptor(
make_tuple(make_merge_transform(make_tuple(K0, N0, H0, W0))), make_tuple(make_merge_transform(make_tuple(K0, N0, H0, W0))),
...@@ -464,15 +527,6 @@ struct GridwiseGemmDlops_km_kn_mn_v3_add ...@@ -464,15 +527,6 @@ struct GridwiseGemmDlops_km_kn_mn_v3_add
const CBlockIdToBlockClusterAdaptor_K_N_H_W& c_blockid_to_k_n_h_w_block_cluster_adaptor, const CBlockIdToBlockClusterAdaptor_K_N_H_W& c_blockid_to_k_n_h_w_block_cluster_adaptor,
integral_constant<bool, HasMainE0BlockLoop>) integral_constant<bool, HasMainE0BlockLoop>)
{ {
// constexpr auto a_e0_e1_k0_k1_e2_grid_desc = AGridDesc_E0_E1_K0_K1_E2{};
// constexpr auto b_e0_e1_n_h0_h1_h2_w0_w1_w2_e2_grid_desc =
// BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2{};
// constexpr auto d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc =
// DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx{};
// constexpr auto c_blockid_to_k_n_h_w_block_cluster_adaptor =
// CBlockIdToBlockClusterAdaptor_K_N_H_W{};
const auto bias_k0_k1_grid_desc = const auto bias_k0_k1_grid_desc =
MakeBiasK0K1GridDescriptor(d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc); MakeBiasK0K1GridDescriptor(d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc);
......
composable_kernel/include/utility/config.hpp
View file @ 4eb9a7a4
...@@ -90,8 +90,9 @@ ...@@ -90,8 +90,9 @@
#endif #endif
// pass tensor descriptor by value or void* // pass tensor descriptor by value or void*
#define CK_EXPERIMENTAL_PASS_TENSOR_DESCRIPTOR_BY_VALUE 1 #define CK_EXPERIMENTAL_PASS_TENSOR_DESCRIPTOR_BY_VALUE 0
#define CK_EXPERIMENTAL_PASS_TENSOR_DESCRIPTOR_BY_VOID_POINTER 0 #define CK_EXPERIMENTAL_PASS_TENSOR_DESCRIPTOR_BY_VOID_POINTER 0
#define CK_EXPERIMENTAL_STATIC_TENSOR_DESCRIPTOR 1
// merge transformation use magic number division // merge transformation use magic number division
#define CK_EXPERIMENTAL_MERGE_USE_MAGIC_DIVISION 0 #define CK_EXPERIMENTAL_MERGE_USE_MAGIC_DIVISION 0
......
host/driver_offline/include/driver_convolution_maxpool_forward_implicit_gemm_v5r1_dlops_nc0hwc1_kc0yxc1_nk0hwk1.hpp
View file @ 4eb9a7a4
...@@ -183,12 +183,12 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -183,12 +183,12 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{})); make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}));
// max tensor // max tensor
const auto d_k_n_hopx2_wopx2_grid_desc = transform_tensor_descriptor( const auto d_k_n_hx_wx_grid_desc = transform_tensor_descriptor(
make_naive_tensor_descriptor_packed(make_tuple(N, K0, Hx, Wx, K1)), make_naive_tensor_descriptor_packed(make_tuple(N, K0, Hx, Wx, K1)),
make_tuple(make_merge_transform(make_tuple(K0, K1)), make_tuple(make_merge_transform(make_tuple(K0, K1)),
make_pass_through_transform(N), make_pass_through_transform(N),
make_pad_transform(Hx, I0, Number<OutRightPadH/2>{}), make_pad_transform(Hx, I0, Number<OutRightPadH / 2>{}),
make_pad_transform(Wx, I0, Number<OutRightPadW/2>{})), make_pad_transform(Wx, I0, Number<OutRightPadW / 2>{})),
make_tuple(Sequence<1, 4>{}, Sequence<0>{}, Sequence<2>{}, Sequence<3>{}), make_tuple(Sequence<1, 4>{}, Sequence<0>{}, Sequence<2>{}, Sequence<3>{}),
make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{})); make_tuple(Sequence<0>{}, Sequence<1>{}, Sequence<2>{}, Sequence<3>{}));
...@@ -289,10 +289,6 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -289,10 +289,6 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
Sequence<0, 0, 0, 0, 0, 0, 0, 0, 0>{})); Sequence<0, 0, 0, 0, 0, 0, 0, 0, 0>{}));
// clang-format on // clang-format on
static_assert(a_e0_e1_k_e2_grid_desc.IsKnownAtCompileTime(), "");
static_assert(b_e0_e1_n_ho_wo_e2_grid_desc.IsKnownAtCompileTime(), "");
static_assert(d_k_n_hopx2_wopx2_grid_desc.IsKnownAtCompileTime(), "");
static_assert(c_k_n_hop_wop_grid_desc.IsKnownAtCompileTime(), "");
// GEMM // GEMM
using GridwiseGemm = GridwiseGemmDlops_km_kn_mn_v3_add< using GridwiseGemm = GridwiseGemmDlops_km_kn_mn_v3_add<
...@@ -304,7 +300,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -304,7 +300,7 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
decltype(a_e0_e1_k_e2_grid_desc), decltype(a_e0_e1_k_e2_grid_desc),
decltype(b_e0_e1_n_ho_wo_e2_grid_desc), decltype(b_e0_e1_n_ho_wo_e2_grid_desc),
decltype(c_k_n_hop_wop_grid_desc), decltype(c_k_n_hop_wop_grid_desc),
decltype(d_k_n_hopx2_wopx2_grid_desc), decltype(d_k_n_hx_wx_grid_desc),
E1, E1,
E2, E2,
K2, K2,
...@@ -348,14 +344,13 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -348,14 +344,13 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
const auto c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc = const auto c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc =
GridwiseGemm::MakeCK0K1NH0H1H2W0W1W2GridDescriptor(c_k_n_hop_wop_grid_desc); GridwiseGemm::MakeCK0K1NH0H1H2W0W1W2GridDescriptor(c_k_n_hop_wop_grid_desc);
const auto d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc = const auto d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc =
GridwiseGemm::MakeDK0K1NH0H1HxW0W1WxGridDescriptor(d_k_n_hopx2_wopx2_grid_desc); GridwiseGemm::MakeDK0K1NH0H1HxW0W1WxGridDescriptor(d_k_n_hx_wx_grid_desc);
using AGridDesc_E0_E1_K0_K1_E2 = decltype(a_e0_e1_k0_k1_e2_grid_desc); using AGridDesc_E0_E1_K0_K1_E2 = decltype(a_e0_e1_k0_k1_e2_grid_desc);
using BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2 = using BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2 =
decltype(b_e0_e1_n_h0_h1_h2_w0_w1_w2_e2_grid_desc); decltype(b_e0_e1_n_h0_h1_h2_w0_w1_w2_e2_grid_desc);
using CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2 = decltype(c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc); using CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2 = decltype(c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc);
using DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx = using DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx = decltype(d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc);
decltype(d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc);
const auto grid_size = (K / KPerBlock) * (Hop / HoPerBlock) * (Wop / WoPerBlock) * N; const auto grid_size = (K / KPerBlock) * (Hop / HoPerBlock) * (Wop / WoPerBlock) * N;
...@@ -375,16 +370,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -375,16 +370,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
if(has_main_e0_block_loop) if(has_main_e0_block_loop)
{ {
const auto kernel = kernel_gemm_dlops_v2_add< const auto kernel =
GridwiseGemm, kernel_gemm_dlops_v2_add<GridwiseGemm,
FloatAB, FloatAB,
FloatC, FloatC,
remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>, remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>,
remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>, remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>,
remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>, remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>,
remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>, remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>,
remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>, remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>,
true>; true>;
ave_time = launch_and_time_kernel(kernel, ave_time = launch_and_time_kernel(kernel,
nrepeat, nrepeat,
...@@ -404,16 +399,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -404,16 +399,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
} }
else else
{ {
const auto kernel = kernel_gemm_dlops_v2_add< const auto kernel =
GridwiseGemm, kernel_gemm_dlops_v2_add<GridwiseGemm,
FloatAB, FloatAB,
FloatC, FloatC,
remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>, remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>,
remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>, remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>,
remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>, remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>,
remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>, remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>,
remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>, remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>,
false>; false>;
ave_time = launch_and_time_kernel(kernel, ave_time = launch_and_time_kernel(kernel,
nrepeat, nrepeat,
...@@ -456,16 +451,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -456,16 +451,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
if(has_main_e0_block_loop) if(has_main_e0_block_loop)
{ {
const auto kernel = kernel_gemm_dlops_v2_add< const auto kernel =
GridwiseGemm, kernel_gemm_dlops_v2_add<GridwiseGemm,
FloatAB, FloatAB,
FloatC, FloatC,
remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>, remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>,
remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>, remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>,
remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>, remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>,
remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>, remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>,
remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>, remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>,
true>; true>;
ave_time = launch_and_time_kernel( ave_time = launch_and_time_kernel(
kernel, kernel,
...@@ -492,16 +487,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -492,16 +487,16 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
else else
{ {
const auto kernel = kernel_gemm_dlops_v2_add< const auto kernel =
GridwiseGemm, kernel_gemm_dlops_v2_add<GridwiseGemm,
FloatAB, FloatAB,
FloatC, FloatC,
remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>, remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>,
remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>, remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>,
remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>, remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>,
remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>, remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>,
remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>, remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>,
false>; false>;
ave_time = launch_and_time_kernel( ave_time = launch_and_time_kernel(
kernel, kernel,
...@@ -525,7 +520,36 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0 ...@@ -525,7 +520,36 @@ struct DriverDynamicConvolutionForwardImplicitGemmDlops_v5r1_nc0hwc1_kc0yxc1_nk0
cast_pointer_to_constant_address_space( cast_pointer_to_constant_address_space(
c_blockid_to_k_n_h_w_block_cluster_adaptor_dev_buf.GetDeviceBuffer())); c_blockid_to_k_n_h_w_block_cluster_adaptor_dev_buf.GetDeviceBuffer()));
} }
#elif CK_EXPERIMENTAL_STATIC_TENSOR_DESCRIPTOR
{
static_assert(a_e0_e1_k_e2_grid_desc.IsKnownAtCompileTime(), "");
static_assert(b_e0_e1_n_h0_h1_h2_w0_w1_w2_e2_grid_desc.IsKnownAtCompileTime(), "");
static_assert(d_k0_k1_n_h0_h1_hx_w0_w1_wx_grid_desc.IsKnownAtCompileTime(), "");
static_assert(c_k0_k1_n_h0_h1_h2_w0_w1_w2_grid_desc.IsKnownAtCompileTime(), "");
static_assert(c_blockid_to_k_n_h_w_block_cluster_adaptor.IsKnownAtCompileTime(), "");
const auto kernel =
kernel_gemm_dlops_v2_add<GridwiseGemm,
FloatAB,
FloatC,
remove_reference_t<AGridDesc_E0_E1_K0_K1_E2>,
remove_reference_t<BGridDesc_E0_E1_N_H0_H1_H2_W0_W1_W2_E2>,
remove_reference_t<CGridDesc_K0_K1_N_H0_H1_H2_W0_W1_W2>,
remove_reference_t<DGridDesc_K0_K1_N_H0_H1_Hx_W0_W1_Wx>,
remove_reference_t<CBlockIdToBlockClusterAdaptor_K_N_H_W>,
has_main_e0_block_loop>;
ave_time = launch_and_time_kernel(kernel,
nrepeat,
dim3(grid_size),
dim3(BlockSize),
0,
p_a_grid,
p_b_grid,
p_bias_grid,
p_c_grid,
p_d_grid);
}
#endif #endif
return ave_time; return ave_time;
} }
......
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