Merge branch 'develop' of...

Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/composable_kernel into add_int8_wmma_example_instance

example/36_sparse_embedding/sparse_embedding3_forward_layernorm.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include <iostream>
 #include <numeric>

example/37_batched_gemm_add_add_relu_gemm_add/CMakeLists.txt

-add_example_executable(example_batched_gemm_add_add_relu_gemm_add_xdl_fp16 batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp)
+if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
+    add_example_executable(example_batched_gemm_add_add_relu_gemm_add_xdl_fp16 batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp)
+endif()

example/37_batched_gemm_add_add_relu_gemm_add/batched_gemm_add_add_relu_gemm_add_xdl_fp16.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 /*
 Computes C_m_o = Relu(A0[m, k] * B0[n, k] + D00[m, n] + D01[mn]) * B1[n, o] + D1[m, o]
@@ -173,6 +173,8 @@ using DeviceGemmInstance =
         8,
         8,
         true,
+        9,           // D0sTransferSrcVectorDim
+        4,           // D0sTransferSrcScalaerPerVector
         S<8, 32, 1>, // B1BlockTransfer
         S<0, 2, 1>,
         S<0, 2, 1>,
@@ -189,7 +191,7 @@ int main(int argc, char* argv[])
 {
     bool do_verification = true;
     int init_method      = 1;
-    bool time_kernel     = false;
+    bool time_kernel     = true;
 
     // GEMM shape
     ck::index_t M              = 1024;

example/38_grouped_conv_bwd_data_multiple_d/CMakeLists.txt

-if(GPU_TARGETS MATCHES "gfx908" OR GPU_TARGETS MATCHES "gfx90a" OR GPU_TARGETS MATCHES "gfx940")
+if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
+list(APPEND gpu_list gfx908 gfx90a gfx940 gfx941 gfx942)
+set(target 0)
+foreach(gpu IN LISTS GPU_TARGETS)
+ if(gpu IN_LIST gpu_list AND target EQUAL 0)
    add_custom_target(example_grouped_conv_bwd_data)
    add_example_executable(example_grouped_conv_bwd_data_fp16 grouped_conv_bwd_data_fp16.cpp)
    add_example_executable(example_grouped_conv_bwd_data_bias_relu_fp16 grouped_conv_bwd_data_bias_relu_fp16.cpp)
 
    add_dependencies(example_grouped_conv_bwd_data example_grouped_conv_bwd_data_fp16)
    add_dependencies(example_grouped_conv_bwd_data example_grouped_conv_bwd_data_bias_relu_fp16)
-endif()
\ No newline at end of file
+   set(target 1)
+ endif()
+endforeach()
+endif()

example/38_grouped_conv_bwd_data_multiple_d/common.hpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

example/38_grouped_conv_bwd_data_multiple_d/grouped_conv_bwd_data_bias_relu_fp16.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
 

example/38_grouped_conv_bwd_data_multiple_d/grouped_conv_bwd_data_fp16.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
 

example/38_grouped_conv_bwd_data_multiple_d/run_grouped_conv_bwd_data_bias_relu_example.inc

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 bool run_conv_bwd_data_bias_relu(const ExecutionConfig& config,
                                  const ck::utils::conv::ConvParam& conv_params,

example/38_grouped_conv_bwd_data_multiple_d/run_grouped_conv_bwd_data_example.inc

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 bool run_conv_bwd_data(const ExecutionConfig& config,
                        const ck::utils::conv::ConvParam& conv_params,

example/39_permute/CMakeLists.txt

-add_custom_target(example_permute)
+if(DTYPES MATCHES "fp16" OR NOT DEFINED DTYPES)
+    add_custom_target(example_permute)
 
-add_example_executable(example_permute_1xHxW_fp16 permute_1xHxW_fp16.cpp)
-add_example_executable(example_permute_NxHxW_fp16 permute_NxHxW_fp16.cpp)
-add_example_executable(example_permute_HxWx4_fp16 permute_HxWx4_fp16.cpp)
+    add_example_executable(example_permute_1xHxW_fp16 permute_1xHxW_fp16.cpp)
+    add_example_executable(example_permute_NxHxW_fp16 permute_NxHxW_fp16.cpp)
+    add_example_executable(example_permute_HxWx4_fp16 permute_HxWx4_fp16.cpp)
 
-add_dependencies(example_permute example_permute_1xHxW_fp16)
-add_dependencies(example_permute example_permute_NxHxW_fp16)
-add_dependencies(example_permute example_permute_HxWx4_fp16)
+    add_dependencies(example_permute example_permute_1xHxW_fp16)
+    add_dependencies(example_permute example_permute_NxHxW_fp16)
+    add_dependencies(example_permute example_permute_HxWx4_fp16)
+endif()

example/39_permute/common.hpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

example/39_permute/permute_1xHxW_fp16.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
 

example/39_permute/permute_HxWx4_fp16.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
 

example/39_permute/permute_NxHxW_fp16.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
 

example/39_permute/run_permute_bundle_example.inc

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

example/39_permute/run_permute_element_example.inc

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

example/40_conv2d_fwd_quantization/CMakeLists.txt

-if(GPU_TARGETS MATCHES "gfx908" OR GPU_TARGETS MATCHES "gfx90a" OR GPU_TARGETS MATCHES "gfx940")
+if(DTYPES MATCHES "int8" OR NOT DEFINED DTYPES)
+list(APPEND gpu_list gfx908 gfx90a gfx940 gfx941 gfx942)
+set(target 0)
+foreach(gpu IN LISTS GPU_TARGETS)
+ if(gpu IN_LIST gpu_list AND target EQUAL 0)
    add_example_executable(example_conv2d_fwd_xdl_perlayer_quantization_int8 conv2d_fwd_xdl_perlayer_quantization_int8.cpp)
    add_example_executable(example_conv2d_fwd_xdl_perchannel_quantization_int8 conv2d_fwd_xdl_perchannel_quantization_int8.cpp)
    add_example_executable(example_conv2d_fwd_xdl_bias_relu_perlayer_quantization_int8 conv2d_fwd_xdl_bias_relu_perlayer_quantization_int8.cpp)
    add_example_executable(example_conv2d_fwd_xdl_bias_relu_perchannel_quantization_int8 conv2d_fwd_xdl_bias_relu_perchannel_quantization_int8.cpp)
-endif()
-# Conv perlayer quantization
-add_example_executable(example_conv2d_fwd_dl_perlayer_quantization_int8 conv2d_fwd_dl_perlayer_quantization_int8.cpp)
+   set(target 1)
+ endif()
+endforeach()
 
-# Conv perchannel quantization
-add_example_executable(example_conv2d_fwd_dl_perchannel_quantization_int8 conv2d_fwd_dl_perchannel_quantization_int8.cpp)
-
-# Conv + bias + relu perlayer quantization
-add_example_executable(example_conv2d_fwd_dl_bias_relu_perlayer_quantization_int8 conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp)
-
-# Conv + bias + relu perchannel quantization
-add_example_executable(example_conv2d_fwd_dl_bias_relu_perchannel_quantization_int8 conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp)
-
-# Conv + bias + tanh perlayer quantization
-add_example_executable(example_conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8 conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp)
-
-# Conv + bias + tanh perchannel quantization
-add_example_executable(example_conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8 conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp)
\ No newline at end of file
+  if(DL_KERNELS)
+  # Conv perlayer quantization
+  add_example_executable(example_conv2d_fwd_dl_perlayer_quantization_int8 conv2d_fwd_dl_perlayer_quantization_int8.cpp)
+  # Conv perchannel quantization
+  add_example_executable(example_conv2d_fwd_dl_perchannel_quantization_int8 conv2d_fwd_dl_perchannel_quantization_int8.cpp)
+  # Conv + bias + relu perlayer quantization
+  add_example_executable(example_conv2d_fwd_dl_bias_relu_perlayer_quantization_int8 conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp)
+  # Conv + bias + relu perchannel quantization
+  add_example_executable(example_conv2d_fwd_dl_bias_relu_perchannel_quantization_int8 conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp)
+  # Conv + bias + tanh perlayer quantization
+  add_example_executable(example_conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8 conv2d_fwd_dl_bias_tanh_perlayer_quantization_int8.cpp)
+  # Conv + bias + tanh perchannel quantization
+  add_example_executable(example_conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8 conv2d_fwd_dl_bias_tanh_perchannel_quantization_int8.cpp)
+  endif()
+endif()
\ No newline at end of file

example/40_conv2d_fwd_quantization/common.hpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #pragma once
 

example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perchannel_quantization_int8.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp"
 
 using InDataType           = int8_t;
 using WeiDataType          = int8_t;

example/40_conv2d_fwd_quantization/conv2d_fwd_dl_bias_relu_perlayer_quantization_int8.cpp

 // SPDX-License-Identifier: MIT
-// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
 
 #include "common.hpp"
-#include "ck/tensor_operation/gpu/device/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_dl_multiple_d_nhwc_kyxc_nhwk.hpp"
 
 using InDataType   = int8_t;
 using WeiDataType  = int8_t;