COMPMID-1836: Remove CLGEMMTranspose1xWKernel and replace with CLGEMMReshapeRHSMatrixKernel

Change-Id: Ic5a4f32657a155380684dcd4b44fbb608ef40cb4
Reviewed-on: https://review.mlplatform.org/418
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>

6 files changed, 11 insertions, 15 deletions

diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.h
index 82dcd93ce6..616c269b0d 100644
--- a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.h
+++ b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.h
@@ -58,7 +58,7 @@ public:
      * @param[in]  input0                    Input tensor containing the interleaved Matrix A. Data type supported: QASYMM8
      * @param[in]  input1                    Input tensor containing the transposed1xW Matrix B. Data type supported: same as @p input0
      * @param[out] output                    Output tensor to store the result of matrix multiplication. Data type supported: S32
-     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMTranspose1xWKernel
+     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMReshapeRHSMatrixKernel
      * @param[in]  reshape_info              (Optional) GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
      */
     void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, bool is_interleaved_transposed = true, const GEMMReshapeInfo &reshape_info = GEMMReshapeInfo());
@@ -67,7 +67,7 @@ public:
      * @param[in] input0                    Input tensor info containing the interleaved Matrix A. Data type supported: QASYMM8
      * @param[in] input1                    Input tensor info containing the transposed Matrix B. Data type supported: same as @p input0
      * @param[in] output                    Output tensor info to store the result of matrix multiplication. Data type supported: S32
-     * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMTranspose1xWKernel
+     * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMReshapeRHSMatrixKernel
      * @param[in] reshape_info              GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
      *
      * @return a status
diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
index f61c330de6..ce37787862 100644
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
+++ b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
@@ -32,7 +32,7 @@ class ICLTensor;
 
 /** OpenCL kernel to multiply two input matrices "A" and "B" . All elements of the output matrix will be multiplied by alpha
  *
- * @note If the input tensors @p input0 and @p input1 have been reshaped respectively with @ref CLGEMMInterleave4x4Kernel" and @ref CLGEMMTranspose1xWKernel,
+ * @note If the input tensors @p input0 and @p input1 have been reshaped respectively with @ref CLGEMMInterleave4x4Kernel" and @ref CLGEMMReshapeRHSMatrixKernel,
  *       the flag @p is_interleaved_transposed must be set to true
  *
  * @attention The second input tensor must have at least 2 dimensions (matrix)
@@ -57,7 +57,7 @@ public:
      * @param[in]  input1                    Input tensor containing the Matrix B. Data type supported: same as @p input0
      * @param[out] output                    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
      * @param[in]  alpha                     Weight of the matrix product
-     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMTranspose1xWKernel
+     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMReshapeRHSMatrixKernel
      * @param[in]  reshape_info              (Optional) GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
      * @param[in]  fp_mixed_precision        (Optional) Use wider accumulators (32 bit instead of 16 for FP16) to improve accuracy
      *
@@ -70,7 +70,7 @@ public:
      * @param[in] input1                    Input tensor containing the Matrix B. Data type supported: same as @p input0
      * @param[in] output                    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
      * @param[in] alpha                     Weight of the matrix product
-     * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMTranspose1xWKernel
+     * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMInterleave4x4Kernel and @ref CLGEMMReshapeRHSMatrixKernel
      * @param[in] reshape_info              GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
      * @param[in] gpu_target                GPU Target
      * @param[in] fp_mixed_precision        (Optional) Use wider accumulators (32 bit instead of 16 for FP16) to improve accuracy
diff --git a/arm_compute/core/Types.h b/arm_compute/core/Types.h
index 6ef9878a95..02001a2438 100644
--- a/arm_compute/core/Types.h
+++ b/arm_compute/core/Types.h
@@ -1652,8 +1652,8 @@ private:
  * The matrix A can only be reshaped through @ref CLGEMMInterleave4x4Kernel or  @ref NEGEMMInterleave4x4Kernel or  @ref GCGEMMInterleave4x4Kernel
  * Note: Optionally just for @ref CLGEMMInterleave4x4Kernel is it possible to set mult_interleave4x4_height, the multiplication factor for the height of the 4x4 interleaved block
  *
- * The matrix B can only be reshaped through @ref CLGEMMTranspose1xWKernel or  @ref NEGEMMTranspose1xWKernel or  @ref GCGEMMTranspose1xWKernel
- * Note: Optionally just for @ref CLGEMMTranspose1xWKernel is it possible to set mult_transpose1xW_width, the multiplication factor for the width of the 1xW transposed block
+ * The matrix B can only be reshaped through @ref CLGEMMReshapeRHSMatrixKernel or  @ref NEGEMMTranspose1xWKernel or  @ref GCGEMMTranspose1xWKernel
+ * Note: Optionally just for @ref CLGEMMReshapeRHSMatrixKernel is it possible to set mult_transpose1xW_width, the multiplication factor for the width of the 1xW transposed block
  *
  */
 class GEMMReshapeInfo final
diff --git a/arm_compute/runtime/CL/functions/CLGEMM.h b/arm_compute/runtime/CL/functions/CLGEMM.h
index 7d47194e56..c4accde23d 100644
--- a/arm_compute/runtime/CL/functions/CLGEMM.h
+++ b/arm_compute/runtime/CL/functions/CLGEMM.h
@@ -30,7 +30,6 @@
 #include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMReshapeLHSMatrixKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMTranspose1xWKernel.h"
 #include "arm_compute/runtime/CL/CLMemoryGroup.h"
 #include "arm_compute/runtime/CL/CLTensor.h"
 #include "arm_compute/runtime/IFunction.h"
@@ -44,8 +43,7 @@ class ICLTensor;
  *
  *  -# @ref CLGEMMInterleave4x4Kernel (only if the reshaped GEMM is selected by the heuristic model and the GPU target is NOT Mali-G76)
  *  -# @ref CLGEMMReshapeLHSMatrixKernel (only if the reshaped GEMM is selected by the heuristic model and the GPU target IS Mali-G76)
- *  -# @ref CLGEMMTranspose1xWKernel (only if the reshaped GEMM is selected by the heuristic model and the GPU target is NOT Mali-G76)
- *  -# @ref CLGEMMReshapeRHSMatrixKernel (only if the reshaped GEMM is selected by the heuristic model and the GPU target IS Mali-G76)
+ *  -# @ref CLGEMMReshapeRHSMatrixKernel (only if the reshaped GEMM is selected by the heuristic model)
  *  -# @ref CLGEMMMatrixMultiplyKernel (if GPU target is NOT G76 or if the reshaped GEMM is NOT selected)
  *  -# @ref CLGEMMMatrixMultiplyReshapedKernel (only if the reshaped GEMM is selected by the heuristic model and the GPU target IS Mali-G76)
  *  -# @ref CLGEMMMatrixAdditionKernel (if c != nullptr and beta != 0.0)
@@ -108,7 +106,6 @@ public:
 private:
     CLMemoryGroup                      _memory_group;
     CLGEMMInterleave4x4Kernel          _interleave_kernel; // TODO - COMPMID-1835: Remove this kernel and use CLGEMMReshapeLHSMatrixKernel
-    CLGEMMTranspose1xWKernel           _transpose_kernel;  // TODO - COMPMID-1836: Remove this kernel and use CLGEMMReshapeRHSMatrixKernel
     CLGEMMMatrixMultiplyKernel         _mm_kernel;
     CLGEMMMatrixAdditionKernel         _ma_kernel;
     CLGEMMReshapeLHSMatrixKernel       _reshape_lhs_kernel;
diff --git a/arm_compute/runtime/CL/functions/CLGEMMConvolutionLayer.h b/arm_compute/runtime/CL/functions/CLGEMMConvolutionLayer.h
index 1468b156eb..d7694a8328 100644
--- a/arm_compute/runtime/CL/functions/CLGEMMConvolutionLayer.h
+++ b/arm_compute/runtime/CL/functions/CLGEMMConvolutionLayer.h
@@ -30,7 +30,6 @@
 #include "arm_compute/core/CL/kernels/CLElementwiseOperationKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMInterleave4x4Kernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMTranspose1xWKernel.h"
 #include "arm_compute/core/CL/kernels/CLIm2ColKernel.h"
 #include "arm_compute/core/CL/kernels/CLWeightsReshapeKernel.h"
 #include "arm_compute/core/Types.h"
diff --git a/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h b/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h
index 82f307a773..141354e723 100644
--- a/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h
+++ b/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h
@@ -29,7 +29,7 @@
 #include "arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpReductionKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMTranspose1xWKernel.h"
+#include "arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h"
 #include "arm_compute/runtime/CL/CLMemoryGroup.h"
 #include "arm_compute/runtime/CL/CLTensor.h"
 #include "arm_compute/runtime/IFunction.h"
@@ -42,7 +42,7 @@ class ICLTensor;
 /** Basic function to execute GEMMLowpMatrixMultiplyCore on OpenCL. This function calls the following OpenCL kernels:
  *
  *  -# @ref CLGEMMInterleave4x4Kernel  (if the output tensor is a matrix)
- *  -# @ref CLGEMMTranspose1xWKernel  (if the output tensor is a matrix)
+ *  -# @ref CLGEMMReshapeRHSMatrixKernel  (if the output tensor is a matrix)
  *  -# @ref CLGEMMLowpMatrixMultiplyKernel
  *  -# @ref CLGEMMLowpMatrixAReductionKernel (if the offset of matrix B is not 0)
  *  -# @ref CLGEMMLowpMatrixBReductionKernel (if the offset of matrix A is not 0)
@@ -102,7 +102,7 @@ private:
     CLMemoryGroup                                 _memory_group;
     CLGEMMLowpMatrixMultiplyKernel                _mm_kernel;
     CLGEMMInterleave4x4Kernel                     _mtx_a_reshape_kernel;
-    CLGEMMTranspose1xWKernel                      _mtx_b_reshape_kernel;
+    CLGEMMReshapeRHSMatrixKernel                  _mtx_b_reshape_kernel;
     CLGEMMLowpMatrixAReductionKernel              _mtx_a_reduction_kernel;
     CLGEMMLowpMatrixBReductionKernel              _mtx_b_reduction_kernel;
     CLGEMMLowpOffsetContributionKernel            _offset_contribution_kernel;