Batched Gemm with C Permute (#305)

* init commit

* add c_permute

* add mnk padding

* fixed comments

* Fixed comments
Co-authored-by: Chao Liu <chao.liu2@amd.com>

example/24_batched_gemm_c_permute/CMakeLists.txt

+add_example_executable(example_batched_gemm_c_permute_xdl_fp16 batched_gemm_c_permute_xdl_fp16.cpp)
+

example/24_batched_gemm_c_permute/batched_gemm_c_permute_xdl_fp16.cpp

+#include <iostream>
+#include <numeric>
+#include <initializer_list>
+#include <cstdlib>
+
+#include "ck/ck.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/device_batched_gemm_c_permute_xdl.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+
+#include "ck/library/utility/check_err.hpp"
+#include "ck/library/host_tensor/device_memory.hpp"
+#include "ck/library/host_tensor/host_tensor.hpp"
+#include "ck/library/host_tensor/host_tensor_generator.hpp"
+#include "ck/library/reference_tensor_operation/cpu/reference_batched_gemm.hpp"
+
+template <ck::index_t... Is>
+using S = ck::Sequence<Is...>;
+
+using F16 = ck::half_t;
+using F32 = float;
+
+using Row = ck::tensor_layout::gemm::RowMajor;
+using Col = ck::tensor_layout::gemm::ColumnMajor;
+
+using PassThrough = ck::tensor_operation::element_wise::PassThrough;
+
+using ADataType   = ck::half_t;
+using BDataType   = ck::half_t;
+using CDataType   = ck::half_t;
+using AccDataType = float;
+
+using ALayout = ck::tensor_layout::gemm::RowMajor;
+using BLayout = ck::tensor_layout::gemm::ColumnMajor;
+using CLayout = ck::tensor_layout::gemm::RowMajor;
+
+using AElementOp = ck::tensor_operation::element_wise::PassThrough;
+using BElementOp = ck::tensor_operation::element_wise::PassThrough;
+using CElementOp = ck::tensor_operation::element_wise::PassThrough;
+
+// static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
+// static constexpr auto MNPadding = ck::tensor_operation::device::GemmSpecialization::MNPadding;
+static constexpr auto MNKPadding = ck::tensor_operation::device::GemmSpecialization::MNKPadding;
+
+// clang-format off
+using DeviceGemmInstance = ck::tensor_operation::device::DeviceBatchedGemmCPermuteXdl
+//######| ALayout| BLayout| AData| BData| CData| AccData|           A|           B|           C|          GEMM|      Num| Block|  MPer|  NPer|  KPer| AK1| BK1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds|     CShuffle|    CShuffle| CBlockTransferClusterLengths|   CBlockTransfer|
+//######|        |        |  Type|  Type|  Type|    Type| Elementwise| Elementwise| Elementwise|Spacialization| Prefetch|  Size| Block| Block| Block|    |    |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN|  MXdlPerWave| NXdlPerWave|         _MBlock_MWaveMPerXdl|  ScalarPerVector|
+//######|        |        |      |      |      |        |   Operation|   Operation|   Operation|              |         |      |      |      |      |    |    |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |   PerShuffle|  PerShuffle|         _NBlock_NWaveNPerXdl|    _NWaveNPerXdl|
+//######|        |        |      |      |      |        |            |            |            |              |         |      |      |      |      |    |    |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |             |            |                             |                 |
+//      <     Row,     Col,   F16,   F16,   F16,     F32, PassThrough, PassThrough, PassThrough,     MNPadding,        1,   256,   256,   128,    32,   8,   8,   32,   32,    4,    2,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,             2,              8,              8,      true,            1,           1,               S<1, 32, 1, 8>,                8>;
+        <     Row,     Col,   F16,   F16,   F16,     F32, PassThrough, PassThrough, PassThrough,     MNKPadding,        1,   256,   128,    64,    32,   8,   8,   32,   32,    2,    1,     S<4, 64, 1>,     S<1, 0, 2>,     S<1, 0, 2>,              2,              8,              8,      true,     S<4, 32, 1>,     S<1, 0, 2>,     S<1, 0, 2>,             2,              8,              8,      true,            1,           1,               S<1, 32, 1, 8>,                8>;
+// clang-format on
+
+using ReferenceBatchedGemmInstance = ck::tensor_operation::host::
+    ReferenceBatchedGemm<ADataType, BDataType, CDataType, AElementOp, BElementOp, CElementOp>;
+
+int main(int argc, char* argv[])
+{
+    bool do_verification = true;
+    int init_method      = 1;
+    bool time_kernel     = false;
+
+    const int M = 88;
+    const int N = 64;
+    const int K = 88;
+
+    const int stride_A = K;
+    const int stride_B = K;
+
+    const int G0 = 1024;
+    const int G1 = 10;
+
+    const int batch_count = G0 * G1;
+
+    // output layout - [G0, M, G1, N]
+    const int stride_G0 = M * G1 * N;
+    const int stride_G1 = N;
+    const int stride_M  = G1 * N;
+    const int stride_N  = 1;
+
+    if(argc == 4)
+    {
+        do_verification = std::stoi(argv[1]);
+        init_method     = std::stoi(argv[2]);
+        time_kernel     = std::stoi(argv[3]);
+    }
+    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=n0, 1=yes)\n");
+        exit(0);
+    }
+
+    // GEMM shape
+    ck::tensor_operation::device::BatchedGemmCPermuteDesc batched_gemm_c_permute_desc{
+        G0, G1, M, N, stride_G0, stride_G1, stride_M, stride_N};
+
+    auto f_host_tensor_descriptor = [](std::size_t batch_count_,
+                                       std::size_t row,
+                                       std::size_t col,
+                                       std::size_t stride,
+                                       auto layout) {
+        if(std::is_same<decltype(layout), ck::tensor_layout::gemm::RowMajor>::value)
+        {
+            return HostTensorDescriptor(std::vector<std::size_t>({batch_count_, row, col}),
+                                        std::vector<std::size_t>({row * stride, stride, 1}));
+        }
+        else
+        {
+            return HostTensorDescriptor(std::vector<std::size_t>({batch_count_, row, col}),
+                                        std::vector<std::size_t>({col * stride, 1, stride}));
+        }
+    };
+
+    Tensor<ADataType> a_g_m_k(f_host_tensor_descriptor(batch_count, M, K, stride_A, ALayout{}));
+    Tensor<BDataType> b_g_k_n(f_host_tensor_descriptor(batch_count, K, N, stride_B, BLayout{}));
+
+    auto f_host_c_tensor_descriptor = [](std::size_t G0_,
+                                         std::size_t G1_,
+                                         std::size_t M_,
+                                         std::size_t N_,
+                                         std::size_t stride_G0_,
+                                         std::size_t stride_G1_,
+                                         std::size_t stride_M_,
+                                         std::size_t stride_N_) {
+        return HostTensorDescriptor(
+            std::vector<std::size_t>({G0_, G1_, M_, N_}),
+            std::vector<std::size_t>({stride_G0_, stride_G1_, stride_M_, stride_N_}));
+    };
+
+    Tensor<CDataType> c_g0_g1_m_n_host_result(
+        f_host_c_tensor_descriptor(G0, G1, M, N, stride_G0, stride_G1, stride_M, stride_N));
+
+    Tensor<CDataType> c_g0_g1_m_n_device_result(
+        f_host_c_tensor_descriptor(G0, G1, M, N, stride_G0, stride_G1, stride_M, stride_N));
+
+    std::cout << "a_g_m_k: " << a_g_m_k.mDesc << std::endl;
+    std::cout << "b_g_k_n: " << b_g_k_n.mDesc << std::endl;
+    std::cout << "c_g0_g1_m_n: " << c_g0_g1_m_n_host_result.mDesc << std::endl;
+
+    switch(init_method)
+    {
+    case 0: break;
+    case 1:
+        a_g_m_k.GenerateTensorValue(GeneratorTensor_2<ADataType>{-5, 5});
+        b_g_k_n.GenerateTensorValue(GeneratorTensor_2<BDataType>{-5, 5});
+        break;
+    default:
+        a_g_m_k.GenerateTensorValue(GeneratorTensor_3<ADataType>{0.0, 1.0});
+        b_g_k_n.GenerateTensorValue(GeneratorTensor_3<BDataType>{-0.5, 0.5});
+        break;
+    }
+
+    DeviceMem a_device_buf(sizeof(ADataType) * a_g_m_k.mDesc.GetElementSpace());
+    DeviceMem b_device_buf(sizeof(BDataType) * b_g_k_n.mDesc.GetElementSpace());
+    DeviceMem c_device_buf(sizeof(CDataType) * c_g0_g1_m_n_device_result.mDesc.GetElementSpace());
+
+    a_device_buf.ToDevice(a_g_m_k.mData.data());
+    b_device_buf.ToDevice(b_g_k_n.mData.data());
+
+    auto a_element_op = AElementOp{};
+    auto b_element_op = BElementOp{};
+    auto c_element_op = CElementOp{};
+
+    auto gemm    = DeviceGemmInstance{};
+    auto invoker = gemm.MakeInvoker();
+
+    // do GEMM
+    auto argument = gemm.MakeArgument(static_cast<ADataType*>(a_device_buf.GetDeviceBuffer()),
+                                      static_cast<BDataType*>(b_device_buf.GetDeviceBuffer()),
+                                      static_cast<CDataType*>(c_device_buf.GetDeviceBuffer()),
+                                      M,
+                                      N,
+                                      K,
+                                      stride_A,
+                                      stride_B,
+                                      batched_gemm_c_permute_desc,
+                                      a_element_op,
+                                      b_element_op,
+                                      c_element_op,
+                                      batch_count);
+
+    if(!gemm.IsSupportedArgument(argument))
+    {
+        throw std::runtime_error(
+            "wrong! device_gemm with the specified compilation parameters does "
+            "not support this GEMM problem");
+    }
+
+    float ave_time = invoker.Run(argument, StreamConfig{nullptr, time_kernel});
+
+    std::size_t flop      = std::size_t(2) * batch_count * M * N * K;
+    std::size_t num_btype = sizeof(ADataType) * batch_count * M * K +
+                            sizeof(BDataType) * batch_count * K * N +
+                            sizeof(CDataType) * batch_count * 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, "
+              << gemm.GetTypeString() << std::endl;
+
+    bool pass = true;
+
+    if(do_verification)
+    {
+        c_device_buf.FromDevice(c_g0_g1_m_n_device_result.mData.data());
+
+        auto ref_batched_gemm = ReferenceBatchedGemmInstance{};
+        auto ref_invoker      = ref_batched_gemm.MakeInvoker();
+
+        Tensor<CDataType> c_g_m_n_host_result = HostTensorDescriptor(
+            std::vector<std::size_t>({batch_count, M, N}), std::vector<std::size_t>({M * N, N, 1}));
+
+        auto ref_argument = ref_batched_gemm.MakeArgument(
+            a_g_m_k, b_g_k_n, c_g_m_n_host_result, a_element_op, b_element_op, c_element_op);
+
+        ref_invoker.Run(ref_argument);
+
+        for(int g0 = 0; g0 < G0; g0++)
+        {
+            for(int g1 = 0; g1 < G1; g1++)
+            {
+                for(int m = 0; m < M; m++)
+                {
+                    for(int n = 0; n < N; n++)
+                    {
+                        int g                                 = g0 * G1 + g1;
+                        c_g0_g1_m_n_host_result(g0, g1, m, n) = c_g_m_n_host_result(g, m, n);
+                    }
+                }
+            }
+        }
+
+        pass = ck::utils::check_err(c_g0_g1_m_n_host_result.mData,
+                                    c_g0_g1_m_n_device_result.mData,
+                                    "Error: Incorrect results c");
+    }
+
+    return pass ? 0 : 1;
+}

example/CMakeLists.txt

@@ -42,4 +42,5 @@ add_subdirectory(20_convnd_bwd_weight_xdl)
 add_subdirectory(21_gemm_layernorm)
 add_subdirectory(22_cgemm)
 add_subdirectory(23_softmax)
+add_subdirectory(24_batched_gemm_c_permute)
 add_subdirectory(25_gemm_bias_c_permute)

include/ck/tensor_operation/gpu/device/device_batched_gemm_c_permute.hpp

+#pragma once
+#include <iostream>
+#include <vector>
+
+#include "device_base.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+
+struct BatchedGemmCPermuteDesc
+{
+    ck::index_t G0_, G1_, M_, N_;
+    ck::index_t stride_G0_, stride_G1_, stride_M_, stride_N_;
+};
+
+template <typename AElementwiseOperation,
+          typename BElementwiseOperation,
+          typename CElementwiseOperation>
+struct DeviceBatchedGemmCPermute : public BaseOperator
+{
+    virtual std::unique_ptr<BaseArgument>
+    MakeArgumentPointer(const void* p_a,
+                        const void* p_b,
+                        void* p_c,
+                        index_t M,
+                        index_t N,
+                        index_t K,
+                        index_t stride_A,
+                        index_t stride_B,
+                        BatchedGemmCPermuteDesc batched_gemm_c_permute_desc,
+                        AElementwiseOperation a_element_op,
+                        BElementwiseOperation b_element_op,
+                        CElementwiseOperation c_element_op,
+                        ck::index_t BatchCount) = 0;
+
+    virtual std::unique_ptr<BaseInvoker> MakeInvokerPointer() = 0;
+};
+
+template <typename AElementwiseOperation,
+          typename BElementwiseOperation,
+          typename CElementwiseOperation>
+using DeviceBatchedGemmCPermutePtr = std::unique_ptr<
+    DeviceBatchedGemmCPermute<AElementwiseOperation, BElementwiseOperation, CElementwiseOperation>>;
+
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck