From d325aee11872a810354f2ead8a102d0e0ff8389d Mon Sep 17 00:00:00 2001
From: Eunju Yang <ej.yang@samsung.com>
Date: Fri, 29 Nov 2024 15:17:01 +0900
Subject: [PATCH] [ GPU/OpenCL ] change transpose_cl to inherit LayerImplCl

- This commit updates transpose_cl.cpp/h to inherit LayerImplCl.
- This commit implements registerClKernels() of transpose_cl layer.
- This commit update cl_context.cpp (applying transpose_cl's update)
- This is the last commit to complete #2723.
- This can close #2723.

Self evaluation:

Build test: [X]Passed [ ]Failed [ ]Skipped
Run test: [X]Passed [ ]Failed [ ]Skipped

Signed-off-by: Eunju Yang <ej.yang@samsung.com>
---
 nntrainer/cl_context.cpp                  |  9 +++++----
 nntrainer/layers/cl_layers/transpose_cl.h | 22 +++++++++++++---------
 2 files changed, 18 insertions(+), 13 deletions(-)

diff --git a/nntrainer/cl_context.cpp b/nntrainer/cl_context.cpp
index 25c2d66c2b..6696c06d9f 100644
--- a/nntrainer/cl_context.cpp
+++ b/nntrainer/cl_context.cpp
@@ -64,10 +64,11 @@ static void add_default_object(ClContext &cc) {
                        ConcatLayerCl::type, ml::train::LayerType::LAYER_CONCAT);
   }
 
-  // @todo transposlayercl also needs to be updated.
-  cc.registerFactory(nntrainer::createLayer<TransposeLayerCl>,
-                     TransposeLayerCl::type,
-                     ml::train::LayerType::LAYER_TRANSPOSE);
+  if (TransposeLayerCl::registerClKernels()) {
+    cc.registerFactory(nntrainer::createLayer<TransposeLayerCl>,
+                       TransposeLayerCl::type,
+                       ml::train::LayerType::LAYER_TRANSPOSE);
+  }
 }
 
 static void registerer(ClContext &cc) noexcept {
diff --git a/nntrainer/layers/cl_layers/transpose_cl.h b/nntrainer/layers/cl_layers/transpose_cl.h
index af6d3bd301..42bacc3b55 100644
--- a/nntrainer/layers/cl_layers/transpose_cl.h
+++ b/nntrainer/layers/cl_layers/transpose_cl.h
@@ -16,6 +16,7 @@
 
 #include <common_properties.h>
 #include <layer_devel.h>
+#include <layer_impl_cl.h>
 #include <opencl_buffer.h>
 #include <opencl_kernel.h>
 
@@ -25,13 +26,13 @@ namespace nntrainer {
  * @brief A tranpose layer.
  *
  */
-class TransposeLayerCl final : public Layer {
+class TransposeLayerCl final : public LayerImplCl {
 public:
   /**
    * @brief Construct a new transpose layer object
    *
    */
-  TransposeLayerCl() : Layer(), transpose_props(props::Print()) {}
+  TransposeLayerCl() : LayerImplCl(), transpose_props(props::Print()) {}
 
   /**
    * @brief Destroy the transpose layer object
@@ -82,15 +83,18 @@ class TransposeLayerCl final : public Layer {
    */
   void setProperty(const std::vector<std::string> &values) override;
 
-  inline static const std::string type = "transpose";
+  /**
+   * @brief     registerClKernels for transpose_cl
+   * @details   registerClKernels for transpose_cl always returns true
+   * without any specific action for kernel registeration. It only uses
+   * cl_blas_kernels and there is no specific kernels for this. If there are
+   * specific kernels for this, it should be updated to register the kernels .
+   */
+  static bool registerClKernels() { return true; };
 
-  static opencl::Kernel kernel_transpose_axis0;
-  static opencl::Kernel kernel_transpose_fp16_axis0;
-  static opencl::Kernel kernel_transpose_axis1;
-  static opencl::Kernel kernel_transpose_fp16_axis1;
-  static opencl::Kernel kernel_transpose_axis2;
-  static opencl::Kernel kernel_transpose_fp16_axis2;
+  inline static const std::string type = "transpose";
 
+private:
   std::tuple<props::Print> transpose_props; /**< transpose layer properties :
                                             unit - number of output neurons */
 };