Merge remote-tracking branch 'upstream/develop' into gemm-layernorm-4

library/src/tensor_operation_instance/gpu/gemm/device_gemm_xdl_splitk_f32_f32_f32_km_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_splitk.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_splitk.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm/device_gemm_xdl_splitk_f32_f32_f32_km_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_splitk.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_splitk.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm/device_gemm_xdl_splitk_f32_f32_f32_mk_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_splitk.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_splitk.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm/device_gemm_xdl_splitk_f32_f32_f32_mk_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_splitk.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_splitk.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_add_add_fastgelu/device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_km_kn_mn_instance.cpp

-#include <stdlib.h>
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
 
-#include "config.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
-#include "device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -25,9 +28,9 @@ using AddAddFastGelu = ck::tensor_operation::element_wise::AddAddFastGelu;
 
 static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
 
-// e = elementwise((a * b), d)
+// e = elementwise((a * b), d0, d1)
 // outout: e[m, n]
-// input: a[k, m], b[k, n], d[m, n]
+// input: a[k, m], b[k, n], d0[m, n], d1[m, n]
 using device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_km_kn_mn_instances = std::tuple<
     // clang-format off
         //##############################| ALayout| BLayout| ELayout| AData| BData| AccData| CShuffle|  DsData| EData|           A|           B|            CDE|           GEMM| NumGemmK| 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|

library/src/tensor_operation_instance/gpu/gemm_add_add_fastgelu/device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_km_nk_mn_instance.cpp

-#include <stdlib.h>
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
 
-#include "config.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
-#include "device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -25,9 +28,9 @@ using AddAddFastGelu = ck::tensor_operation::element_wise::AddAddFastGelu;
 
 static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
 
-// e = elementwise((a * b), d)
+// e = elementwise((a * b), d0, d1)
 // outout: e[m, n]
-// input: a[k, m], b[n, k], d[m, n]
+// input: a[k, m], b[n, k], d0[m, n], d1[m, n]
 using device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_km_nk_mn_instances = std::tuple<
     // clang-format off
         //##############################| ALayout| BLayout| ELayout| AData| BData| AccData| CShuffle|  DsData| EData|           A|           B|            CDE|           GEMM| NumGemmK| 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|

library/src/tensor_operation_instance/gpu/gemm_add_add_fastgelu/device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_mk_kn_mn_instance.cpp

-#include <stdlib.h>
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
 
-#include "config.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
-#include "device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -25,9 +28,9 @@ using AddAddFastGelu = ck::tensor_operation::element_wise::AddAddFastGelu;
 
 static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
 
-// e = elementwise((a * b), d)
+// e = elementwise((a * b), d0, d1)
 // outout: e[m, n]
-// input: a[m, k], b[k, n], d[m, n]
+// input: a[m, k], b[k, n], d0[m, n], d1[m, n]
 using device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_mk_kn_mn_instances = std::tuple<
     // clang-format off
         //##############################| ALayout| BLayout| CLayout| AData| BData| AccData| CShuffle|  DsData| EData|           A|           B|            CDE|           GEMM| NumGemmK| 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|

library/src/tensor_operation_instance/gpu/gemm_add_add_fastgelu/device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_mk_nk_mn_instance.cpp

-#include <stdlib.h>
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
 
-#include "config.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
-#include "device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/device_gemm_multiple_d_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -25,9 +28,9 @@ using AddAddFastGelu = ck::tensor_operation::element_wise::AddAddFastGelu;
 
 static constexpr auto GemmDefault = ck::tensor_operation::device::GemmSpecialization::Default;
 
-// e = elementwise((a * b), d)
+// e = elementwise((a * b), d0, d1)
 // outout: e[m, n]
-// input: a[m, k], b[n, k], d[m, n]
+// input: a[m, k], b[n, k], d0[m, n], d1[m ,n]
 using device_gemm_add_add_fastgelu_xdl_c_shuffle_f16_f16_f16_mk_nk_mn_instances = std::tuple<
     // clang-format off
         //##############################| ALayout| BLayout| CLayout| AData| BData| AccData| CShuffle|  DsData| EData|           A|           B|            CDE|           GEMM| NumGemmK| 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|

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f16_f16_f16_km_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f16_f16_f16_km_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f16_f16_f16_mk_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f16_f16_f16_mk_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f32_f32_f32_km_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f32_f32_f32_km_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f32_f32_f32_mk_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias2d/device_gemm_xdl_c_shuffle_bias_2d_f32_f32_f32_mk_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_xdl_c_shuffle_bias_2d.hpp"
-#include "element_wise_operation.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#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_gemm_xdl_c_shuffle_bias_2d.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias_add_reduce/CMakeLists.txt

@@ -6,5 +6,5 @@ set(DEVICE_GEMM_REDUCE_INSTANCE_SOURCE
 )
 
 add_instance_library(device_gemm_bias_add_reduce_instance ${DEVICE_GEMM_REDUCE_INSTANCE_SOURCE})
-install(TARGETS device_gemm_bias_add_reduce_instance LIBRARY DESTINATION lib)
+rocm_install(TARGETS device_gemm_bias_add_reduce_instance)
 clang_tidy_check(device_gemm_bias_add_reduce_instance)

library/src/tensor_operation_instance/gpu/gemm_bias_add_reduce/device_gemm_bias_add_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_bias_add_reduce_xdl_cshuffle.hpp"
-#include "element_wise_operation.hpp"
-#include "reduction_operator.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/utility/reduction_operator.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_gemm_bias_add_reduce_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias_add_reduce/device_gemm_bias_add_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_km_nk_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_bias_add_reduce_xdl_cshuffle.hpp"
-#include "element_wise_operation.hpp"
-#include "reduction_operator.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/utility/reduction_operator.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_gemm_bias_add_reduce_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {

library/src/tensor_operation_instance/gpu/gemm_bias_add_reduce/device_gemm_bias_add_reduce_xdl_cshuffle_f16_f16_f16_f32_f32_mk_kn_mn_instance.cpp

-#include <stdlib.h>
-#include "config.hpp"
-#include "device_gemm_bias_add_reduce_xdl_cshuffle.hpp"
-#include "element_wise_operation.hpp"
-#include "reduction_operator.hpp"
-#include "device_operation_instance.hpp"
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/utility/reduction_operator.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_gemm_bias_add_reduce_xdl_cshuffle.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+#include "ck/library/tensor_operation_instance/device_operation_instance.hpp"
 
 namespace ck {
 namespace tensor_operation {