/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/device/device_gemm_bias_e_permute.hpp Source File

/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/device/device_gemm_bias_e_permute.hpp Source File#

Composable Kernel: /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/device/device_gemm_bias_e_permute.hpp Source File

Go to the documentation of this file.

 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
  
 #pragma once
  
 #include <array>
  
 #include "device_base.hpp"
  
 namespace ck {
 namespace tensor_operation {
 namespace device {
  
 struct DEGridDesc_M0_M1_M2_N0_N1
 {
     ck::index_t M0_, M1_, M2_, N0_, N1_;
     ck::index_t stride_M0_, stride_M1_, stride_M2_, stride_N0_, stride_N1_;
 };
  
 // input : A[M, K], B[K, N],
 // input : D[M, N], ...
 // output : E[M, N]
 // C = a_op(A) * b_op(B)
 // E = cde_op(C, D)
 template <typename AElementwiseOperation,
           typename BElementwiseOperation,
           typename CDEElementwiseOperation>
 struct DeviceGemmBiasCPermute : public BaseOperator
 {
     virtual std::unique_ptr<BaseArgument>
     MakeArgumentPointer(const void* p_a,
                         const void* p_b,
                         const void* p_d,
                         void* p_e,
                         ck::index_t M,
                         ck::index_t N,
                         ck::index_t K,
                         ck::index_t StrideA,
                         ck::index_t StrideB,
                         DEGridDesc_M0_M1_M2_N0_N1 d_gride_desc,
                         DEGridDesc_M0_M1_M2_N0_N1 e_gride_desc,
                         AElementwiseOperation a_element_op,
                         BElementwiseOperation b_element_op,
                         CDEElementwiseOperation cde_element_op) = 0;
  
     virtual std::unique_ptr<BaseInvoker> MakeInvokerPointer() = 0;
 };
  
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck