Add example of groupnorm splitK

example/42_groupnorm/CMakeLists.txt

 add_example_executable(example_groupnorm_sigmoid_mul_fp16 groupnorm_sigmoid_mul_fp16.cpp)
+add_example_executable(example_groupnorm_splitk_fp16 groupnorm_splitk_fp16.cpp)
 add_example_executable(example_groupnorm_swish_fp16 groupnorm_swish_fp16.cpp)

example/42_groupnorm/common.hpp

@@ -12,6 +12,7 @@
 #include "ck/ck.hpp"
 #include "ck/utility/reduction_enums.hpp"
 #include "ck/tensor_operation/gpu/device/impl/device_normalization_impl.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_normalization_splitk_impl.hpp"
 #include "ck/tensor_operation/gpu/device/reduction_operator_mapping.hpp"
 
 #include "ck/library/utility/fill.hpp"

example/42_groupnorm/groupnorm_splitk_fp16.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "common.hpp"
+
+constexpr int Rank         = 5;
+constexpr int NumReduceDim = 3;
+
+using XDataType       = ck::half_t;
+using GammaDataType   = ck::half_t;
+using BetaDataType    = ck::half_t;
+using YDataType       = ck::half_t;
+using ComputeDataType = float;
+using YElementOp      = ck::tensor_operation::element_wise::Swish;
+
+using DeviceInstance =
+    ck::tensor_operation::device::DeviceNormalizationSplitKImpl<XDataType,
+                                                                GammaDataType,
+                                                                BetaDataType,
+                                                                ComputeDataType,
+                                                                YDataType,
+                                                                YElementOp,
+                                                                Rank,
+                                                                NumReduceDim,
+                                                                256, // BlockSize
+                                                                1,   // ClusterM
+                                                                256, // ClusterK
+                                                                1,   // SliceM
+                                                                16,  // SliceK
+                                                                1,   // SrcVecDim (0=M, 1=K)
+                                                                2,   // SrcScalarPerVector
+                                                                1,   // GammaVecDim (0=M, 1=K)
+                                                                2,   // GammaScalarPerVector
+                                                                1,   // BetaVecDim (0=M, 1=K)
+                                                                2,   // BetaScalarPerVector
+                                                                2>;  // OutScalarPerVector
+
+#include "run_groupnorm_example.inc"
+
+int main(int argc, char* argv[]) { run_groupnorm_example(argc, argv); }

example/42_groupnorm/run_groupnorm_example.inc

@@ -73,6 +73,10 @@ int run_groupnorm_example(int argc, char* argv[])
         return 1;
     };
 
+    size_t workspace_sz = device_instance.GetWorkSpaceSize(argument_ptr.get());
+    DeviceMem workspace_dev(workspace_sz);
+    device_instance.SetWorkSpacePointer(argument_ptr.get(), workspace_dev.GetDeviceBuffer());
+
     auto invoker_ptr = device_instance.MakeInvokerPointer();
     float ave_time   = invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, true, true});