Hip tensor permute unit test (#1068)

* adding files for F32 example

* adding functioning implementation with scalar multiplication and unary operator support

* added fp 16 type check in unary square

* updating scalar multiplication as an operator

* functioning version with scalar operator

* changing strides for col major

* updated column major implementation

* working column major implementation

* cleaned up comments, rearranged/renamed files

* small edits to 3d transpose profiler

* adding test/profiler/instance files for hipTensor permute unit test

* added more test instances

* cleaned up errors, randomized input tensor, added more instances

* turned off time printouts

* removed conflicting transpose profiler

* rearranged some files

library/include/ck/library/tensor_operation_instance/gpu/permute_scale.hpp

@@ -0,0 +1,77 @@
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#pragma once
+
+#include <cstdlib>
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
+#include "ck/tensor_operation/gpu/device/device_elementwise_scale.hpp"
+#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
+
+#include "ck/library/tensor_operation_instance/device_operation_instance_factory.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_permute_scale_f16_instances(
+    std::vector<std::unique_ptr<DeviceElementwise<ck::Tuple<F16>,
+                                                  ck::Tuple<F16>,
+                                                  PassThrough,
+                                                  element_wise::UnarySquare,
+                                                  Scale,
+                                                  4>>>&);
+
+void add_device_permute_scale_f32_instances(
+    std::vector<std::unique_ptr<DeviceElementwise<ck::Tuple<F32>,
+                                                  ck::Tuple<F32>,
+                                                  PassThrough,
+                                                  element_wise::UnarySquare,
+                                                  Scale,
+                                                  4>>>&);
+
+template <typename InDataTypeTuple,
+          typename OutDataTypeTuple,
+          typename ElementwiseOperation,
+          typename UnaryOperation,
+          typename Scale,
+          index_t NumDim>
+struct DeviceOperationInstanceFactory<
+    ck::tensor_operation::device::DeviceElementwise<InDataTypeTuple,
+                                                    OutDataTypeTuple,
+                                                    ElementwiseOperation,
+                                                    UnaryOperation,
+                                                    Scale,
+                                                    NumDim>>
+{
+    using DeviceOp = DeviceElementwise<InDataTypeTuple,
+                                       OutDataTypeTuple,
+                                       ElementwiseOperation,
+                                       UnaryOperation,
+                                       Scale,
+                                       NumDim>;
+
+    static auto GetInstances()
+    {
+        std::vector<std::unique_ptr<DeviceOp>> op_ptrs;
+        if constexpr(is_same_v<InDataTypeTuple, ck::Tuple<F32>> &&
+                     is_same_v<OutDataTypeTuple, ck::Tuple<F32>>)
+        {
+            add_device_permute_scale_f32_instances(op_ptrs);
+        }
+        else if constexpr(is_same_v<InDataTypeTuple, ck::Tuple<F16>> &&
+                          is_same_v<OutDataTypeTuple, ck::Tuple<F16>>)
+        {
+            add_device_permute_scale_f16_instances(op_ptrs);
+        }
+        return op_ptrs;
+    }
+};
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

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

@@ -0,0 +1,2 @@
+add_instance_library(device_permute_scale_instance 
+	device_permute_scale_instances.cpp)

library/src/tensor_operation_instance/gpu/permute_scale/device_permute_scale_instances.cpp

@@ -0,0 +1,56 @@
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/ck.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_elementwise_scale_impl.hpp"
+#include "ck/utility/data_type.hpp"
+
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+using F16 = ck::half_t;
+using F32 = float;
+
+using Pass    = ck::tensor_operation::element_wise::PassThrough;
+using UnaryOp = ck::tensor_operation::element_wise::UnarySquare;
+using Scale   = ck::tensor_operation::element_wise::Scale;
+
+// clang-format off
+using device_permute_scale_f16_instances =
+    std::tuple <
+        DeviceElementwiseImpl<ck::Tuple<F16>, ck::Tuple<F16>, Pass, UnaryOp, Scale, 4, 1, ck::Sequence<1>, ck::Sequence<1>>,
+        DeviceElementwiseImpl<ck::Tuple<F16>, ck::Tuple<F16>, Pass, UnaryOp, Scale, 4, 8, ck::Sequence<1>, ck::Sequence<1>>,
+        DeviceElementwiseImpl<ck::Tuple<F16>, ck::Tuple<F16>, Pass, UnaryOp, Scale, 4, 4, ck::Sequence<1>, ck::Sequence<1>>,
+        DeviceElementwiseImpl<ck::Tuple<F16>, ck::Tuple<F16>, Pass, UnaryOp, Scale, 4, 2, ck::Sequence<1>, ck::Sequence<1>>
+    >;
+
+using device_permute_scale_f32_instances = std::tuple<
+         DeviceElementwiseImpl<ck::Tuple<F32>, ck::Tuple<F32>, Pass, UnaryOp, Scale, 4, 1, ck::Sequence<1>, ck::Sequence<1>>,
+         DeviceElementwiseImpl<ck::Tuple<F32>, ck::Tuple<F32>, Pass, UnaryOp, Scale, 4, 8, ck::Sequence<1>, ck::Sequence<1>>,
+         DeviceElementwiseImpl<ck::Tuple<F32>, ck::Tuple<F32>, Pass, UnaryOp, Scale, 4, 4, ck::Sequence<1>, ck::Sequence<1>>,
+         DeviceElementwiseImpl<ck::Tuple<F32>, ck::Tuple<F32>, Pass, UnaryOp, Scale, 4, 2, ck::Sequence<1>, ck::Sequence<1>>
+    >;
+// clang-format on
+
+void add_device_permute_scale_f16_instances(
+    std::vector<std::unique_ptr<
+        DeviceElementwise<ck::Tuple<F16>, ck::Tuple<F16>, Pass, UnaryOp, Scale, 4>>>& instances)
+{
+    add_device_operation_instances(instances, device_permute_scale_f16_instances{});
+}
+
+void add_device_permute_scale_f32_instances(
+    std::vector<std::unique_ptr<
+        DeviceElementwise<ck::Tuple<F32>, ck::Tuple<F32>, Pass, UnaryOp, Scale, 4>>>& instances)
+{
+    add_device_operation_instances(instances, device_permute_scale_f32_instances{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/transpose/device_transpose_instances_3d.cpp

@@ -19,22 +19,14 @@ void add_device_transpose_f16_instances(
     std::vector<std::unique_ptr<DeviceElementwise<ck::Tuple<F16>, ck::Tuple<F16>, PassThrough, 5>>>&
         instances)
 {
-#ifdef CK_ENABLE_FP16
     add_device_operation_instances(instances, device_transpose_f16_instances{});
-#else
-    ignore = instances;
-#endif
 }
 
 void add_device_transpose_f32_instances(
     std::vector<std::unique_ptr<DeviceElementwise<ck::Tuple<F32>, ck::Tuple<F32>, PassThrough, 5>>>&
         instances)
 {
-#ifdef CK_ENABLE_FP32
     add_device_operation_instances(instances, device_transpose_f32_instances{});
-#else
-    ignore = instances;
-#endif
 }
 
 } // namespace instance