changes to make it compile

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -62,3 +65,5 @@ using DeviceGemmSplitKPtr = std::unique_ptr<DeviceGemmSplitK<ALayout,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -70,3 +73,5 @@ struct DeviceGroupedContractionMultipleD : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -65,3 +68,5 @@ struct DeviceGroupedConvBwdDataMultipleD : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -48,3 +51,5 @@ struct DeviceGroupedConvBwdWeight : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -53,3 +56,5 @@ struct DeviceGroupedConvFwd : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -63,3 +66,5 @@ struct DeviceGroupedConvFwdMultipleD : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -53,3 +56,5 @@ struct DeviceGroupedGemm : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -73,3 +76,5 @@ struct DeviceGroupedGemmSoftmaxGemmPermute : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 #pragma once
 #include <iostream>
 #include <vector>
@@ -37,3 +40,5 @@ struct DeviceGroupedGemmSplitK : public DeviceGroupedGemm<ALayout,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -56,3 +59,5 @@ using DeviceMultipleReducePtr = std::unique_ptr<DeviceMultipleReduce<Rank,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -60,3 +63,5 @@ using DeviceNormalizationPtr = std::unique_ptr<DeviceNormalization<XDataType,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 
+#ifndef __HIPCC_RTC__
 #include <array>
 #include <memory>
 #include <type_traits>
+#endif
 
 #include "ck/tensor_operation/gpu/device/device_base.hpp"
 
@@ -34,3 +39,5 @@ struct DevicePermute : BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -45,3 +48,5 @@ struct DevicePoolFwd : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -69,3 +72,5 @@ using DeviceReducePtr = std::unique_ptr<DeviceReduce<InDataType,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -70,3 +73,5 @@ using DeviceSoftmaxPtr = std::unique_ptr<DeviceSoftmax<InDataType,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -63,3 +66,5 @@ struct DeviceSplitKContractionMultipleD : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
@@ -28,7 +31,7 @@ enum struct GemmSpecialization
     NKOPadding,
     MNKOPadding,
 };
-
+#ifndef __HIPCC_RTC__
 inline std::string getGemmSpecializationString(const GemmSpecialization& s)
 {
     switch(s)
@@ -52,7 +55,10 @@ inline std::string getGemmSpecializationString(const GemmSpecialization& s)
     default: return "Unrecognized specialization!";
     }
 }
+#endif
 
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_batched_contraction_multiple_d_wmma_cshuffle.hpp

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 
+#ifndef __HIPCC_RTC__
 #include <iostream>
 #include <sstream>
+#include "ck/host_utility/device_prop.hpp"
+#include "ck/host_utility/kernel_launch.hpp"
+#endif
+
 
 #include "ck/utility/common_header.hpp"
 #include "ck/tensor_description/tensor_descriptor.hpp"
@@ -15,8 +23,6 @@
 #include "ck/tensor_operation/gpu/device/tensor_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/matrix_padder.hpp"
 #include "ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_wmma_cshuffle.hpp"
-#include "ck/host_utility/device_prop.hpp"
-#include "ck/host_utility/kernel_launch.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -687,7 +693,7 @@ struct DeviceBatchedContractionMultipleD_Wmma_CShuffle
         // Batch Offset
         ComputePtrOffsetOfStridedBatch compute_ptr_offset_of_batch_;
     };
-
+#ifndef __HIPCC_RTC__
     // Invoker
     struct Invoker : public BaseInvoker
     {
@@ -761,13 +767,14 @@ struct DeviceBatchedContractionMultipleD_Wmma_CShuffle
             return Run(*dynamic_cast<const Argument*>(p_arg), stream_config);
         }
     };
+#endif
 
     static constexpr bool IsValidCompilationParameter()
     {
         // TODO: properly implement this check
         return true;
     }
-
+#ifndef __HIPCC_RTC__
     static bool IsSupportedArgument(const Argument& arg)
     {
         if(ck::get_device_name() == "gfx1100" || ck::get_device_name() == "gfx1101" ||
@@ -869,7 +876,7 @@ struct DeviceBatchedContractionMultipleD_Wmma_CShuffle
     {
         return IsSupportedArgument(*dynamic_cast<const Argument*>(p_arg));
     }
-
+#endif
     static auto
     MakeArgument(const void* p_a,
                  const void* p_b,
@@ -943,7 +950,8 @@ struct DeviceBatchedContractionMultipleD_Wmma_CShuffle
                                           cde_element_op);
     }
 
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
 
     // polymorphic
     std::unique_ptr<BaseInvoker> MakeInvokerPointer() override
@@ -985,8 +993,11 @@ struct DeviceBatchedContractionMultipleD_Wmma_CShuffle
 
         return str.str();
     }
+#endif
 };
 
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_batched_contraction_multiple_d_xdl_cshuffle.hpp

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 
+#ifndef __HIPCC_RTC__
 #include <iostream>
 #include <sstream>
+#include "ck/host_utility/device_prop.hpp"
+#include "ck/host_utility/kernel_launch.hpp"
+#endif
+
 
 #include "ck/utility/common_header.hpp"
 #include "ck/tensor_description/tensor_descriptor.hpp"
@@ -15,8 +23,6 @@
 #include "ck/tensor_operation/gpu/device/tensor_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/matrix_padder.hpp"
 #include "ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_xdl_cshuffle.hpp"
-#include "ck/host_utility/device_prop.hpp"
-#include "ck/host_utility/kernel_launch.hpp"
 
 namespace ck {
 
@@ -759,6 +765,7 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
         ComputePtrOffsetOfStridedBatch compute_ptr_offset_of_batch_;
     };
 
+#ifndef __HIPCC_RTC__
     // Invoker
     struct Invoker : public BaseInvoker
     {
@@ -841,7 +848,6 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
             return Run(*dynamic_cast<const Argument*>(p_arg), stream_config);
         }
     };
-
     static bool IsSupportedArgument(const Argument& arg)
     {
         if(!ck::is_xdl_supported())
@@ -935,7 +941,7 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
     {
         return IsSupportedArgument(*dynamic_cast<const Argument*>(p_arg));
     }
-
+#endif
     static auto
     MakeArgument(const void* p_a,
                  const void* p_b,
@@ -970,7 +976,10 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
                         cde_element_op};
     }
 
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
+
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -1006,7 +1015,7 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
                                           b_element_op,
                                           cde_element_op);
     }
-
+#ifndef __HIPCC_RTC__
     // polymorphic
     std::unique_ptr<BaseInvoker> MakeInvokerPointer() override
     {
@@ -1038,8 +1047,11 @@ struct DeviceBatchedContractionMultipleD_Xdl_CShuffle
 
         return str.str();
     }
+#endif
 };
 
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_batched_gemm_e_permute_xdl.hpp

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 #pragma once
 
+#ifndef __HIPCC_RTC__
 #include <iostream>
 #include <sstream>
+#include "ck/host_utility/device_prop.hpp"
+#include "ck/host_utility/kernel_launch.hpp"
+#endif
+
 
 #include "ck/utility/common_header.hpp"
 #include "ck/tensor_description/tensor_descriptor.hpp"
@@ -11,8 +19,6 @@
 #include "ck/tensor_operation/gpu/device/gemm_specialization.hpp"
 #include "ck/tensor_operation/gpu/device/matrix_padder.hpp"
 #include "ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_xdl_cshuffle.hpp"
-#include "ck/host_utility/device_prop.hpp"
-#include "ck/host_utility/kernel_launch.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -490,6 +496,7 @@ struct DeviceBatchedGemmEPermuteXdl : public DeviceBatchedGemmEPermute<ALayout,
         CDEElementwiseOperation cde_element_op_;
     };
 
+#ifndef __HIPCC_RTC__
     // Invoker
     struct Invoker : public BaseInvoker
     {
@@ -565,13 +572,14 @@ struct DeviceBatchedGemmEPermuteXdl : public DeviceBatchedGemmEPermute<ALayout,
             return Run(*dynamic_cast<const Argument*>(p_arg), stream_config);
         }
     };
+#endif
 
     static constexpr bool IsValidCompilationParameter()
     {
         // TODO: properly implement this check
         return true;
     }
-
+#ifndef __HIPCC_RTC__
     static bool IsSupportedArgument(const Argument& arg)
     {
         if(!ck::is_xdl_supported())
@@ -591,7 +599,7 @@ struct DeviceBatchedGemmEPermuteXdl : public DeviceBatchedGemmEPermute<ALayout,
     {
         return IsSupportedArgument(*dynamic_cast<const Argument*>(p_arg));
     }
-
+#endif
     static auto MakeArgument(const ADataType* p_a,
                              const BDataType* p_b,
                              EDataType* p_e,
@@ -625,7 +633,10 @@ struct DeviceBatchedGemmEPermuteXdl : public DeviceBatchedGemmEPermute<ALayout,
                         cde_element_op};
     }
 
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
+
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -662,6 +673,7 @@ struct DeviceBatchedGemmEPermuteXdl : public DeviceBatchedGemmEPermute<ALayout,
                                           cde_element_op);
     }
 
+#ifndef __HIPCC_RTC__
     // polymorphic
     std::unique_ptr<BaseInvoker> MakeInvokerPointer() override
     {
@@ -685,8 +697,11 @@ struct DeviceBatchedGemmEPermuteXdl : public DeviceBatchedGemmEPermute<ALayout,
 
         return str.str();
     }
+#endif
 };
 
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop