changes to make it compile

include/ck/tensor_operation/gpu/device/impl/device_grouped_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.
 
@@ -846,8 +849,9 @@ struct DeviceGroupedContractionMultipleD_Xdl_CShuffle
                         b_element_op,
                         cde_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -911,3 +915,5 @@ struct DeviceGroupedContractionMultipleD_Xdl_CShuffle
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_data_multiple_d_xdl_cshuffle_v1.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.
 
@@ -993,8 +996,9 @@ struct DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
                         b_element_op,
                         cde_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     std::unique_ptr<BaseArgument> MakeArgumentPointer(
         const void* p_a,                                                 // output image
@@ -1076,3 +1080,5 @@ struct DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_gnwc_gkxc_gnwk_dl.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.
 
@@ -1155,8 +1158,9 @@ struct DeviceGroupedConvBwdWeightGnwcGkxcGnwk_Dl
                         out_element_op,
                         split_k};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     std::unique_ptr<BaseArgument>
     MakeArgumentPointer(const void* p_in_grid,
@@ -1224,3 +1228,5 @@ struct DeviceGroupedConvBwdWeightGnwcGkxcGnwk_Dl
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_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.
 
@@ -1442,8 +1445,9 @@ struct DeviceGroupedConvBwdWeight_Xdl_CShuffle
                         out_element_op,
                         split_k};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     std::unique_ptr<BaseArgument>
     MakeArgumentPointer(const void* p_in_grid,
@@ -1522,3 +1526,5 @@ struct DeviceGroupedConvBwdWeight_Xdl_CShuffle
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.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 <functional>
 #include <iostream>
 #include <iterator>
 #include <numeric>
 #include <sstream>
+#include "ck/host_utility/device_prop.hpp"
+#include "ck/host_utility/kernel_launch.hpp"
+#include "ck/host_utility/io.hpp"
+#endif
+
 
 #include "ck/utility/common_header.hpp"
 #include "ck/tensor_description/tensor_descriptor.hpp"
@@ -19,9 +27,6 @@
 #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_dl_multiple_d.hpp"
-#include "ck/host_utility/device_prop.hpp"
-#include "ck/host_utility/kernel_launch.hpp"
-#include "ck/host_utility/io.hpp"
 
 namespace ck {
 namespace tensor_operation {
@@ -543,7 +548,7 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
                 block_2_ctile_map_ = GridwiseGemm::MakeDefaultBlock2CTileMap(e_grid_desc_m_n_);
             }
         }
-
+#ifndef __HIPCC_RTC__
         void Print() const
         {
             std::cout << "A[K0, M, K1]: " << a_grid_desc_ak0_m_ak1_ << std::endl;
@@ -556,6 +561,7 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
             std::cout << "A[m0, m10, m11, n0, n10, n11]: " << e_grid_desc_m0_m10_m11_n0_n10_n11_
                       << std::endl;
         }
+#endif
 
         //  private:
         // pointers
@@ -704,7 +710,7 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
             return Run(*dynamic_cast<const Argument*>(p_arg), stream_config);
         }
     };
-
+#ifndef __HIPCC_RTC__
     static bool IsSupportedArgument(const Argument& arg)
     {
         namespace ctc = tensor_layout::convolution;
@@ -849,7 +855,7 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
     {
         return IsSupportedArgument(*dynamic_cast<const Argument*>(p_arg));
     }
-
+#endif
     static auto MakeArgument(
         const void* p_a,
         const void* p_b,
@@ -891,9 +897,9 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
                         b_element_op,
                         cde_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
-
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
     std::unique_ptr<BaseArgument> MakeArgumentPointer(
         const void* p_a,
         const void* p_b,
@@ -964,3 +970,5 @@ struct DeviceGroupedConvFwdDlMultipleD_NHWC_KYXC_NHWK
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_nhwc_kyxc_nhwk.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.
 
@@ -595,7 +598,7 @@ struct DeviceGroupedConvFwdDl_NHWC_KYXC_NHWK : public DeviceGroupedConvFwd<NDimS
             return Run(*dynamic_cast<const Argument*>(p_arg), stream_config);
         }
     };
-
+#ifndef __HIPCC_RTC__
     static bool IsSupportedArgument(const Argument& arg)
     {
         namespace ctc = tensor_layout::convolution;
@@ -737,7 +740,7 @@ struct DeviceGroupedConvFwdDl_NHWC_KYXC_NHWK : public DeviceGroupedConvFwd<NDimS
     {
         return IsSupportedArgument(*dynamic_cast<const Argument*>(p_arg));
     }
-
+#endif
     static auto MakeArgument(const void* p_a,
                              const void* p_b,
                              void* p_c,
@@ -773,7 +776,10 @@ struct DeviceGroupedConvFwdDl_NHWC_KYXC_NHWK : public DeviceGroupedConvFwd<NDimS
                         c_element_op};
     }
 
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
+
 
     std::unique_ptr<BaseArgument>
     MakeArgumentPointer(const void* p_a,
@@ -850,3 +856,5 @@ struct DeviceGroupedConvFwdDl_NHWC_KYXC_NHWK : public DeviceGroupedConvFwd<NDimS
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_multiple_d_multiple_r.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.
 
@@ -75,3 +78,5 @@ struct DeviceGroupedConvFwdMultipleDMultipleR : public BaseOperator
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_multiple_d_multiple_r_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.
 
@@ -1025,8 +1028,9 @@ struct DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle
                         qs_element_op,
                         rs_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     std::unique_ptr<BaseArgument> MakeArgumentPointer(
         const void* p_a,
@@ -1115,3 +1119,5 @@ struct DeviceGroupedConvFwdMultipleDMultipleR_Xdl_CShuffle
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_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.
 
@@ -775,8 +778,9 @@ struct DeviceGroupedConvFwdMultipleD_Wmma_CShuffle
                         b_element_op,
                         cde_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     std::unique_ptr<BaseArgument> MakeArgumentPointer(
         const void* p_a,
@@ -852,3 +856,5 @@ struct DeviceGroupedConvFwdMultipleD_Wmma_CShuffle
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_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.
 
@@ -884,8 +887,9 @@ struct DeviceGroupedConvFwdMultipleD_Xdl_CShuffle
                         b_element_op,
                         cde_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     std::unique_ptr<BaseArgument> MakeArgumentPointer(
         const void* p_a,
@@ -964,3 +968,5 @@ struct DeviceGroupedConvFwdMultipleD_Xdl_CShuffle
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 #pragma once
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
@@ -708,8 +711,9 @@ struct DeviceGroupedGemmMultipleD_Dl : public DeviceGroupedGemm<ALayout,
         return Argument{
             p_As, p_Bs, p_Ds, p_Es, gemm_descs, a_element_op, b_element_op, cde_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -764,3 +768,5 @@ struct DeviceGroupedGemmMultipleD_Dl : public DeviceGroupedGemm<ALayout,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_softmax_gemm_permute_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.
 
@@ -587,7 +590,7 @@ struct DeviceGroupedGemmSoftmaxGemmPermute_Xdl_CShuffle
         B1ElementwiseOperation b1_element_op_;
         CElementwiseOperation c_element_op_;
     };
-
+#ifndef __HIPCC_RTC_
     // Invoker
     struct Invoker : public BaseInvoker
     {
@@ -672,13 +675,14 @@ struct DeviceGroupedGemmSoftmaxGemmPermute_Xdl_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::is_xdl_supported())
@@ -779,7 +783,7 @@ struct DeviceGroupedGemmSoftmaxGemmPermute_Xdl_CShuffle
     {
         return IsSupportedArgument(*dynamic_cast<const Argument*>(p_arg));
     }
-
+#endif
     static auto MakeArgument(std::vector<const void*> p_a_vec,
                              std::vector<const void*> p_b_vec,
                              std::vector<const void*> p_b1_vec,
@@ -807,7 +811,9 @@ struct DeviceGroupedGemmSoftmaxGemmPermute_Xdl_CShuffle
                         c_element_op};
     }
 
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -891,3 +897,5 @@ struct DeviceGroupedGemmSoftmaxGemmPermute_Xdl_CShuffle
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

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

+
+#pragma clang diagnostic push
+#pragma clang diagnostic ignored "-Weverything"
 #pragma once
 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
@@ -657,8 +660,9 @@ struct DeviceGroupedGemm_Xdl : public DeviceGroupedGemm<ALayout,
         return Argument{
             p_As, p_Bs, p_Ds, p_Es, gemm_descs, a_element_op, b_element_op, c_element_op};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -719,3 +723,5 @@ struct DeviceGroupedGemm_Xdl : public DeviceGroupedGemm<ALayout,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_xdl_splitk_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.
 
@@ -557,8 +560,9 @@ struct DeviceGroupedGemmXdlSplitKCShuffle : public DeviceGroupedGemmSplitK<ALayo
     {
         return Argument{p_As, p_Bs, p_Es, gemm_descs};
     }
-
-    static auto MakeInvoker() { return Invoker{}; }
+#ifndef __HIPCC_RTC__
+  static auto MakeInvoker() { return Invoker{}; }
+#endif
 
     // polymorphic
     std::unique_ptr<BaseArgument>
@@ -630,3 +634,5 @@ struct DeviceGroupedGemmXdlSplitKCShuffle : public DeviceGroupedGemmSplitK<ALayo
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_multiple_reduce_multiblock.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.
 
@@ -593,3 +596,5 @@ struct DeviceMultipleReduceMultiBlock : public DeviceMultipleReduce<Rank,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_multiple_reduce_threadwise.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.
 
@@ -420,3 +423,5 @@ struct DeviceMultipleReduceThreadWise : public DeviceMultipleReduce<Rank,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_normalization_impl.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.
 
@@ -400,3 +403,5 @@ struct DeviceNormalizationImpl : public DeviceNormalization<XDataType,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_normalization_splitk_impl.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.
 
@@ -656,3 +659,5 @@ struct DeviceNormalizationSplitKImpl : public DeviceNormalization<XDataType,
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_permute_impl.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.
 
@@ -269,14 +272,17 @@ struct DevicePermuteImpl : DevicePermute<NumDim, InDataType, OutDataType, Elemen
     {
         return Argument{std::forward<Args>(args)...};
     }
-
+#ifndef __HIPCC_RTC__
     static std::enable_if_t<std::is_default_constructible_v<Invoker>, Invoker>
     MakeInvoker() noexcept(std::is_nothrow_default_constructible_v<Invoker>)
     {
         return Invoker{};
     }
+#endif
 };
 
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop

include/ck/tensor_operation/gpu/device/impl/device_pool2d_fwd_nhwc_nhwc.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.
 
@@ -111,3 +114,5 @@ struct DevicePool2dFwd_NHWC_NHWC : public DevicePool3dFwd_NDHWC_NDHWC<InDataType
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck
+
+#pragma clang diagnostic pop