ck_gemm.cpp

/*
 * The MIT License (MIT)
 *
 * Copyright (c) 2015-2022 Advanced Micro Devices, Inc. All rights reserved.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */
#include <migraphx/gpu/compiler.hpp>
#include <migraphx/make_op.hpp>
#include <migraphx/gpu/context.hpp>

#include <migraphx/gpu/compile_hip_code_object.hpp>
#include <migraphx/gpu/compile_hip.hpp>
#include <migraphx/ranges.hpp>
#include <migraphx/reduce_dims.hpp>
#include <migraphx/stringutils.hpp>
#include <migraphx/dead_code_elimination.hpp>
#include <migraphx/eliminate_common_subexpression.hpp>
#include <migraphx/module.hpp>
#include <migraphx/pass_manager.hpp>

namespace migraphx {
inline namespace MIGRAPHX_INLINE_NS {
namespace gpu {

// NOLINTNEXTLINE
static const char* const ck_gemm_kernel = R"__migraphx__(
#include <migraphx/kernels/ck_gemm.hpp>
#include <migraphx/kernels/ops.hpp>
#include <migraphx/kernels/integral_constant.hpp>
#include <migraphx/kernels/generic_constant.hpp>
#include <args.hpp>


#include "ck/ck.hpp"
#include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_xdl.hpp"
#include "ck/tensor_operation/gpu/device/device_gemm_xdl_cshuffle.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.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        = F16;
using BDataType        = F16;
using AccDataType      = F32;
using CShuffleDataType = F32;
using CDataType        = F16;

using ALayout = Row;
using BLayout = Col;
using CLayout = Row;

using AElementOp = PassThrough;
using BElementOp = PassThrough;
using CElementOp = PassThrough;

namespace migraphx {

extern "C" {

static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;

using DeviceGemmInstance = ck::tensor_operation::device::DeviceGemmXdl
    // clang-format off
//######|     AData|     BData|     CData|     AccData| ALayout| BLayout| CLayout|           A|           B|           C|          GEMM| Block|  MPer|  NPer| K0Per| K1| MPer| NPer| MXdl| NXdl|  ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds|  BBlockTransfer| BBlockTransfer| BBlockTransfer| BlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CThreadTransfer| CThreadTransfer|
//######|      Type|      Type|      Type|        Type|        |        |        | Elementwise| Elementwise| Elementwise|Spacialization|  Size| Block| Block| Block|   |  XDL|  XDL|  Per|  Per|   ThreadCluster|  ThreadCluster| SrcAccessOrder|   SrcVectorDim|      SrcScalar|      DstScalar| AddExtraM|   ThreadCluster|  ThreadCluster| SrcAccessOrder|  SrcVectorDim|      SrcScalar|      DstScalar| AddExtraN| SrcDstVectorDim|       DstScalar|
//######|          |          |          |            |        |        |        |   Operation|   Operation|   Operation|              |      |      |      |      |   |     |     | Wave| Wave| Lengths_K0_M_K1|   ArrangeOrder|               |               |      PerVector|   PerVector_K1|          | Lengths_K0_N_K1|   ArrangeOrder|               |              |      PerVector|   PerVector_K1|          |                |       PerVector|
//######|          |          |          |            |        |        |        |            |            |            |              |      |      |      |      |   |     |     |     |     |                |               |               |               |               |               |          |                |               |               |              |               |               |          |                |                |
        < ADataType, BDataType, CDataType, AccDataType, ALayout, BLayout, CLayout,  AElementOp,  BElementOp,  CElementOp,   GemmDefault,   256,   256,   128,     4,  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,               7,               1>;
// clang-format on

__global__ void ck_gemm_kernel(void* a_p, void* b_p, void* c_p) 
{
    // GEMM shape
    ck::index_t M = 3840;
    ck::index_t N = 4096;
    ck::index_t K = 4096;

    ck::index_t StrideA = 4096;
    ck::index_t StrideB = 4096;
    ck::index_t StrideC = 4096;

    auto a_element_op = AElementOp{};
    auto b_element_op = BElementOp{};
    auto c_element_op = CElementOp{};

    using AGridDesc_AK0_M_AK1 = decltype(MakeAGridDescriptor_AK0_M_AK1(1, 1, 1));
    using BGridDesc_BK0_N_BK1 = decltype(MakeBGridDescriptor_BK0_N_BK1(1, 1, 1));
    using CGridDesc_M_N       = decltype(MakeCGridDescriptor_M_N(1, 1, 1));

    // GridwiseGemm
    using GridwiseGemm = GridwiseGemm_k0mk1_k0nk1_mn_xdl_cshuffle_v1<
        ADataType, // TODO: distinguish A/B datatype
        AccDataType,
        CShuffleDataType,
        CDataType,
        AElementOp,
        BElementOp,
        CElementOp,
        ck::InMemoryDataOperationEnum::Set,
        AGridDesc_AK0_M_AK1,
        BGridDesc_BK0_N_BK1,
        CGridDesc_M_N,
        NumGemmKPrefetchStage,
        BlockSize,
        MPerBlock,
        NPerBlock,
        KPerBlock,
        AK1,
        BK1,
        MPerXDL,
        NPerXDL,
        MXdlPerWave,
        NXdlPerWave,
        ABlockTransferThreadClusterLengths_AK0_M_AK1,
        ABlockTransferThreadClusterArrangeOrder,
        ABlockTransferSrcAccessOrder,
        ABlockTransferSrcVectorDim,
        ABlockTransferSrcScalarPerVector,
        ABlockTransferDstScalarPerVector_AK1,
        false,
        ABlockLdsExtraM,
        BBlockTransferThreadClusterLengths_BK0_N_BK1,
        BBlockTransferThreadClusterArrangeOrder,
        BBlockTransferSrcAccessOrder,
        BBlockTransferSrcVectorDim,
        BBlockTransferSrcScalarPerVector,
        BBlockTransferDstScalarPerVector_BK1,
        false,
        BBlockLdsExtraN,
        CShuffleMXdlPerWavePerShuffle,
        CShuffleNXdlPerWavePerShuffle,
        CShuffleBlockTransferClusterLengths_MBlock_MPerBlock_NBlock_NPerBlock,
        CShuffleBlockTransferScalarPerVector_NPerBlock,
        LoopSched>;
    
    const auto kernel = kernel_gemm_xdlops_v2r3<
                    GridwiseGemm,
                    ADataType, // TODO: distiguish A/B datatype
                    CDataType,
                    remove_reference_t<DeviceGemmXdl::AGridDesc_K0_M_K1>,
                    remove_reference_t<DeviceGemmXdl::BGridDesc_K0_N_K1>,
                    remove_reference_t<typename GridwiseGemm::CGridDesc_M0_N0_M1_N1_M2_M3_M4_N2>,
                    AElementOp,
                    BElementOp,
                    CElementOp,
                    remove_reference_t<typename GridwiseGemm::DefaultBlock2CTileMap>,
                    true>;
    
    kernel<<<1, 1, 1, 0>>>(p_a, p_b, p_c);
}

}

} // namespace migraphx

)__migraphx__";

struct ck_gemm_compiler : compiler<ck_gemm_compiler>
{
    std::vector<std::string> names() const { return {"ck_gemm"}; }

    operation compile_op(context& ctx, const std::vector<shape>& inputs, const value& v) const
    {
        hip_compile_options options;
        auto out_s = inputs.back();
        options.set_launch_params(v, compute_global_for(ctx, out_s.elements()));
        options.inputs         = inputs;
        options.output         = out_s;
        options.kernel_name    = "ck_gemm_kernel";
        options.virtual_inputs = inputs;

        return compile_hip_code_object(ck_gemm_kernel, options);
    }

    compiler_replace compile(context& ctx, instruction_ref ins, const operation& op) const
    {
        return replace(compile_op(ctx, to_shapes(ins->inputs()), op.to_value()));
    }
};

} // namespace gpu
} // namespace MIGRAPHX_INLINE_NS
} // namespace migraphx