start

example/66_gemm_bias_activation/CMakeLists.txt

+set(GEMM_BIAS_ADD_SOURCES
+    gemm_bias_add_xdl_fp16.cpp
+    gemm_bias_add_fp16.cpp
+)
+add_executable(example_gemm_bias_add_xdl_fp16   ${GEMM_BIAS_ADD_SOURCES})
+target_link_libraries(example_gemm_bias_add_xdl_fp16 PRIVATE utility)

example/66_gemm_bias_activation/gemm_bias_add.hpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#pragma once
+
+#include "ck/ck.hpp"
+#include "ck/stream_config.hpp"
+
+struct GemmBiasAddArgs
+{
+    const void* mat_a;
+    const void* mat_b;
+    const void* mat_bias;
+    void* mat_c;
+    ck::index_t M;
+    ck::index_t N;
+    ck::index_t K;
+};
+
+float gemm_bias_add_fp16(const GemmBiasAddArgs& args, const StreamConfig& config);

example/66_gemm_bias_activation/gemm_bias_add_fp16.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "gemm_bias_add.hpp"
+
+#include "ck/utility/blkgemmpipe_scheduler.hpp"
+#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
+#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_xdl_cshuffle_v3.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/tensor_operation/gpu/element/unary_element_wise_operation.hpp"
+
+using F16 = ck::half_t;
+using FP8 = ck::f8_t;
+using F32 = float;
+
+using Row = ck::tensor_layout::gemm::RowMajor;
+using Col = ck::tensor_layout::gemm::ColumnMajor;
+
+using AccDataType      = F32;
+using CShuffleDataType = F32;
+
+using ALayout  = Row;
+using BLayout  = Col;
+using D0Layout = Row;
+using DsLayout = ck::Tuple<D0Layout>;
+using CLayout  = Row;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+using Add         = ck::tensor_operation::element_wise::Add;
+
+using AElementOp   = PassThrough;
+using BElementOp   = PassThrough;
+using CDEElementOp = Add;
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+static constexpr auto GemmSpec = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
+
+// clang-format off
+template <typename ADataType,   typename BDataType, typename DsDataType,  typename CDataType>
+using DeviceOpInstance_64_16_16_64 = ck::tensor_operation::device::DeviceGemmMultiD_Xdl_CShuffle_V3<
+        Row,      Col, DsLayout, CLayout, ADataType, BDataType,
+        DsDataType, CDataType, AccDataType, CShuffleDataType,
+        AElementOp,  BElementOp, CDEElementOp,       GemmSpec,
+        64,
+        16,   16,    64,
+        8,    8,
+        16,   16,
+        1,    1,
+        S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,
+        2,    8,     8,    0,
+        S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,
+        2,    8,     8,    0,
+        1,    1,
+        S<1, 16, 1, 4>,      S<4, 4>,  ck::BlockGemmPipelineScheduler::Interwave, ck::BlockGemmPipelineVersion::v1, F16>;
+
+template <typename ADataType,   typename BDataType, typename DsDataType,  typename CDataType>
+using DeviceOpInstance_default = ck::tensor_operation::device::DeviceGemmMultiD_Xdl_CShuffle_V3<
+        Row,      Col, DsLayout, CLayout, ADataType, BDataType,
+        DsDataType, CDataType, AccDataType, CShuffleDataType,
+        AElementOp,  BElementOp, CDEElementOp,       GemmSpec,
+        64,
+        16,   16,    64,
+        8,    8,
+        16,   16,
+        1,    1,
+        S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,
+        2,    1,     1,    0,
+        S<8,  8, 1>,     S<1, 0, 2>,    S<1, 0, 2>,
+        2,    1,     1,    0,
+        1,    1,
+        S<1, 16, 1, 4>,      S<1, 1>,  ck::BlockGemmPipelineScheduler::Interwave, ck::BlockGemmPipelineVersion::v1, F16>;
+
+// clang-format on
+
+float gemm_bias_add_fp16(const GemmBiasAddArgs& args, const StreamConfig& config)
+{
+    using ADataType  = ck::half_t;
+    using BDataType  = ck::half_t;
+    using CDataType  = ck::half_t;
+    using D0DataType = F16;
+    using DsDataType = ck::Tuple<D0DataType>;
+
+    if(ck::EnvIsEnabled(CK_ENV(CK_LOGGING)))
+    {
+        std::cout << "gemm_bias_add_fp16: {"
+                  << "mat_a: " << args.mat_a << ",  mat_b: " << args.mat_b
+                  << ", mat_bias: " << args.mat_bias << ",  mat_c: " << args.mat_c
+                  << ", M: " << args.M << ", N: " << args.N << ", K: " << args.K << "}"
+                  << std::endl;
+    }
+
+    auto a_element_op   = AElementOp{};
+    auto b_element_op   = BElementOp{};
+    auto cde_element_op = CDEElementOp{};
+
+    ck::index_t StrideA = args.K;
+    ck::index_t StrideB = args.K;
+    ck::index_t StrideD = 0;
+    ck::index_t StrideC = args.N;
+
+    constexpr ck::index_t NumDTensor = DsDataType::Size();
+
+    float ave_time = 0;
+    auto Run       = [&](auto& gemm) {
+        auto argument = gemm.MakeArgument(args.mat_a,
+                                          args.mat_b,
+                                          std::array<const void*, NumDTensor>{args.mat_bias},
+                                          args.mat_c,
+                                          args.M,
+                                          args.N,
+                                          args.K,
+                                          StrideA,
+                                          StrideB,
+                                          std::array<ck::index_t, NumDTensor>{StrideD},
+                                          StrideC,
+                                          a_element_op,
+                                          b_element_op,
+                                          cde_element_op);
+
+        if(!gemm.IsSupportedArgument(argument))
+        {
+            return false;
+        }
+
+        auto invoker = gemm.MakeInvoker();
+        ave_time     = invoker.Run(argument, config);
+
+        return true;
+    };
+
+    auto gemm = DeviceOpInstance_64_16_16_64<ADataType, BDataType, DsDataType, CDataType>{};
+    if(!Run(gemm))
+    {
+        auto gemm_def = DeviceOpInstance_default<ADataType, BDataType, DsDataType, CDataType>{};
+        Run(gemm_def);
+    }
+
+    return ave_time;
+}

example/66_gemm_bias_activation/gemm_bias_add_xdl_fp16.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
+
+#include <iostream>
+#include <numeric>
+#include <initializer_list>
+#include <cstdlib>
+
+#include "gemm_bias_add.hpp"
+
+#include "ck/library/utility/device_memory.hpp"
+#include "ck/library/utility/host_tensor.hpp"
+#include "ck/library/utility/host_tensor_generator.hpp"
+#include "ck/library/utility/literals.hpp"
+#include "ck/library/reference_tensor_operation/cpu/reference_gemm.hpp"
+#include "ck/library/utility/check_err.hpp"
+#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
+
+using F16 = ck::half_t;
+using FP8 = ck::f8_t;
+using F32 = float;
+
+using A0DataType       = F16;
+using B0DataType       = F16;
+using AccDataType      = F32;
+using CShuffleDataType = F32;
+using D0DataType       = F16;
+using DsDataType       = ck::Tuple<D0DataType>;
+using EDataType        = F16;
+
+using Row = ck::tensor_layout::gemm::RowMajor;
+using Col = ck::tensor_layout::gemm::ColumnMajor;
+
+using A0Layout = Row;
+using B0Layout = Col;
+using D0Layout = Row;
+using DsLayout = ck::Tuple<D0Layout>;
+using ELayout  = Row;
+
+void RunUnfusedTest(const std::vector<ck::half_t>& mat_A,
+                    const std::vector<ck::half_t>& mat_B,
+                    const std::vector<ck::half_t>& mat_C,
+                    std::vector<ck::half_t>& mat_D,
+                    int K,
+                    int M,
+                    int N)
+{
+    for(int m = 0; m < M; m++)
+    {
+        std::vector<float> tmp;
+        for(int n = 0; n < N; n++)
+        {
+            float psum = 0.f;
+            for(int k = 0; k < K; k++)
+            {
+                float areg = float(mat_A[m * K + k]);
+                float breg = float(mat_B[n * K + k]);
+                psum += areg * breg;
+            }
+            psum += ck::type_convert<float>(mat_C[n]);
+            mat_D[m * N + n] = ck::type_convert<ck::half_t>(psum);
+        }
+    }
+}
+
+int main(int argc, char* argv[])
+{
+    bool do_verification = true;
+    int init_method      = 1;
+    bool time_kernel     = true;
+
+    // GEMM shape
+    ck::index_t M = 512;
+    ck::index_t N = 1024;
+    ck::index_t K = 256;
+
+    ck::index_t StrideA = K;
+    ck::index_t StrideB = K;
+    ck::index_t StrideD = 0;
+    ck::index_t StrideE = N;
+
+    if(argc == 1)
+    {
+        // use default case
+    }
+    else if(argc == 4)
+    {
+        do_verification = std::stoi(argv[1]);
+        init_method     = std::stoi(argv[2]);
+        time_kernel     = std::stoi(argv[3]);
+    }
+    else if(argc == 11)
+    {
+        do_verification = std::stoi(argv[1]);
+        init_method     = std::stoi(argv[2]);
+        time_kernel     = std::stoi(argv[3]);
+
+        M = std::stoi(argv[4]);
+        N = std::stoi(argv[5]);
+        K = std::stoi(argv[6]);
+
+        StrideA = std::stoi(argv[7]);
+        StrideB = std::stoi(argv[8]);
+        StrideD = std::stoi(argv[9]);
+        StrideE = std::stoi(argv[10]);
+    }
+    else
+    {
+        printf("arg1: verification (0=no, 1=yes)\n");
+        printf("arg2: initialization (0=no init, 1=integer value, 2=decimal value)\n");
+        printf("arg3: time kernel (0=no, 1=yes)\n");
+        printf("arg4 to 9: M (256x), N(128x), K(32x), StrideA, StrideB, StrideD, StrideE\n");
+        exit(0);
+    }
+
+    auto f_host_tensor_descriptor =
+        [](std::size_t row, std::size_t col, std::size_t stride, auto layout) {
+            using namespace ck::literals;
+
+            if(std::is_same<decltype(layout), ck::tensor_layout::gemm::RowMajor>::value)
+            {
+                return HostTensorDescriptor({row, col}, {stride, 1_uz});
+            }
+            else
+            {
+                return HostTensorDescriptor({row, col}, {1_uz, stride});
+            }
+        };
+
+    Tensor<A0DataType> a0_m_k(f_host_tensor_descriptor(M, K, StrideA, A0Layout{}));
+    Tensor<B0DataType> b0_k_n(f_host_tensor_descriptor(K, N, StrideB, B0Layout{}));
+    Tensor<D0DataType> d0_m_n(f_host_tensor_descriptor(M, N, StrideD, D0Layout{}));
+    Tensor<EDataType> e_m_n_host_result(f_host_tensor_descriptor(M, N, StrideE, ELayout{}));
+    Tensor<EDataType> e_m_n_device_result(f_host_tensor_descriptor(M, N, StrideE, ELayout{}));
+
+    std::cout << "a0_m_k: " << a0_m_k.mDesc << std::endl;
+    std::cout << "b0_k_n: " << b0_k_n.mDesc << std::endl;
+    std::cout << "e_m_n: " << e_m_n_host_result.mDesc << std::endl;
+
+    switch(init_method)
+    {
+    case 0: break;
+    case 1:
+        a0_m_k.GenerateTensorValue(GeneratorTensor_3<A0DataType>{-0.5, 0.5});
+        b0_k_n.GenerateTensorValue(GeneratorTensor_3<B0DataType>{-0.5, 0.5});
+        d0_m_n.GenerateTensorValue(GeneratorTensor_3<D0DataType>{-0.5, 0.5});
+        break;
+    default:
+        a0_m_k.GenerateTensorValue(GeneratorTensor_3<A0DataType>{0.0, 1.0});
+        b0_k_n.GenerateTensorValue(GeneratorTensor_3<B0DataType>{-0.5, 0.5});
+        d0_m_n.GenerateTensorValue(GeneratorTensor_3<D0DataType>{-0.5, 0.5});
+    }
+
+    DeviceMem a0_device_buf(sizeof(A0DataType) * a0_m_k.mDesc.GetElementSpaceSize());
+    DeviceMem b0_device_buf(sizeof(B0DataType) * b0_k_n.mDesc.GetElementSpaceSize());
+    DeviceMem d0_device_buf(sizeof(D0DataType) * d0_m_n.mDesc.GetElementSpaceSize());
+    DeviceMem e_device_buf(sizeof(EDataType) * e_m_n_device_result.mDesc.GetElementSpaceSize());
+
+    a0_device_buf.ToDevice(a0_m_k.mData.data());
+    b0_device_buf.ToDevice(b0_k_n.mData.data());
+    d0_device_buf.ToDevice(d0_m_n.mData.data());
+    e_device_buf.ToDevice(e_m_n_device_result.mData.data());
+
+    GemmBiasAddArgs gemm_args{a0_device_buf.GetDeviceBuffer(),
+                              b0_device_buf.GetDeviceBuffer(),
+                              d0_device_buf.GetDeviceBuffer(),
+                              e_device_buf.GetDeviceBuffer(),
+                              M,
+                              N,
+                              K};
+
+    float ave_time = gemm_bias_add_fp16(gemm_args, StreamConfig{nullptr, time_kernel, 20, 50});
+    // float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel, 20, 50});
+
+    std::size_t flop = std::size_t(2) * M * N * K;
+    std::size_t num_btype =
+        sizeof(A0DataType) * M * K + sizeof(B0DataType) * K * N + sizeof(EDataType) * M * N;
+
+    float tflops = static_cast<float>(flop) / 1.E9 / ave_time;
+
+    float gb_per_sec = num_btype / 1.E6 / ave_time;
+
+    std::cout << "Perf: " << ave_time << " ms, " << tflops << " TFlops, " << gb_per_sec << " GB/s"
+              << std::endl;
+
+    e_device_buf.FromDevice(e_m_n_device_result.mData.data());
+
+    if(do_verification)
+    {
+
+        RunUnfusedTest(a0_m_k.mData, b0_k_n.mData, d0_m_n.mData, e_m_n_host_result.mData, K, M, N);
+
+        e_device_buf.FromDevice(e_m_n_device_result.mData.data());
+
+        return ck::utils::check_err(e_m_n_device_result, e_m_n_host_result) ? 0 : 1;
+    }
+
+    return 0;
+}