Issue COMPMID-1835: Remove CLGEMMInterleave4x4Kernel and replace with CLGEMMReshapeLHSMatrixKernel

Change-Id: Id6a1bd78f9b1698b64a004e4adebc41002b15745
Reviewed-on: https://review.mlplatform.org/496
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>

3 files changed, 19 insertions, 22 deletions

diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
index ce37787862..797bda86cf 100644
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
+++ b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017-2018 ARM Limited.
+ * Copyright (c) 2017-2019 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -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 CLGEMMReshapeRHSMatrixKernel,
+ * @note If the input tensors @p input0 and @p input1 have been reshaped respectively with @ref CLGEMMReshapeLHSMatrixKernel" 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 CLGEMMReshapeRHSMatrixKernel
+     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMReshapeLHSMatrixKernel 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 CLGEMMReshapeRHSMatrixKernel
+     * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMReshapeLHSMatrixKernel 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/runtime/CL/functions/CLGEMM.h b/arm_compute/runtime/CL/functions/CLGEMM.h
index c4accde23d..624df33ef6 100644
--- a/arm_compute/runtime/CL/functions/CLGEMM.h
+++ b/arm_compute/runtime/CL/functions/CLGEMM.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2016-2018 ARM Limited.
+ * Copyright (c) 2016-2019 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -24,7 +24,6 @@
 #ifndef __ARM_COMPUTE_CLGEMM_H__
 #define __ARM_COMPUTE_CLGEMM_H__
 
-#include "arm_compute/core/CL/kernels/CLGEMMInterleave4x4Kernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMMatrixAdditionKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.h"
@@ -41,8 +40,7 @@ class ICLTensor;
 
 /** Basic function to execute GEMM on OpenCL. This function calls the following OpenCL kernels:
  *
- *  -# @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 CLGEMMReshapeLHSMatrixKernel (only if the reshaped GEMM is selected by the heuristic model)
  *  -# @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)
@@ -86,13 +84,13 @@ public:
     void configure(const ICLTensor *a, const ICLTensor *b, const ICLTensor *c, ICLTensor *output, float alpha, float beta, const GEMMInfo &gemm_info = GEMMInfo());
     /** Static function to check if given info will lead to a valid configuration of @ref CLGEMM.
      *
-     * @param[in]  a         First input tensor info  (Matrix or Vector A). Data types supported: F16/F32
-     * @param[in]  b         Second input tensor info (Matrix B). Data type supported: same as @p a.
-     * @param[in]  c         Third input tensor info  (Matrix C). It can be a nullptr if just the multiplication between @p a and @p b is needed. Data type supported: same as @p a.
-     * @param[out] output    Output tensor info. Data type supported: same as @p a
-     * @param[in]  alpha     Weight of the matrix product
-     * @param[in]  beta      Weight of matrix C
-     * @param[in]  gemm_info (Optional) Specifies if the matrix A and/or matrix B have been reshaped and
+     * @param[in] a         First input tensor info  (Matrix or Vector A). Data types supported: F16/F32
+     * @param[in] b         Second input tensor info (Matrix B). Data type supported: same as @p a.
+     * @param[in] c         Third input tensor info  (Matrix C). It can be a nullptr if just the multiplication between @p a and @p b is needed. Data type supported: same as @p a.
+     * @param[in] output    Output tensor info. Data type supported: same as @p a
+     * @param[in] alpha     Weight of the matrix product
+     * @param[in] beta      Weight of matrix C
+     * @param[in] gemm_info (Optional) Specifies if the matrix A and/or matrix B have been reshaped and
      *                       if the reshape of matrix B should happen only for the first run
      *
      * @return a status
@@ -105,7 +103,6 @@ public:
 
 private:
     CLMemoryGroup                      _memory_group;
-    CLGEMMInterleave4x4Kernel          _interleave_kernel; // TODO - COMPMID-1835: Remove this kernel and use CLGEMMReshapeLHSMatrixKernel
     CLGEMMMatrixMultiplyKernel         _mm_kernel;
     CLGEMMMatrixAdditionKernel         _ma_kernel;
     CLGEMMReshapeLHSMatrixKernel       _reshape_lhs_kernel;
@@ -118,8 +115,8 @@ private:
     bool                               _run_addition;
     bool                               _reshape_b_only_on_first_run;
     bool                               _is_prepared;
-    bool                               _is_G76_path; // TODO: To be removed once completed COMPMID-1835 and COMPMID-1836
+    bool                               _is_G76_path;
 };
-}
+} // namespace arm_compute
 
 #endif /* __ARM_COMPUTE_CLGEMM_H__ */
diff --git a/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h b/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h
index 141354e723..72d91070f8 100644
--- a/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h
+++ b/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017-2018 ARM Limited.
+ * Copyright (c) 2017-2019 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -24,11 +24,11 @@
 #ifndef __ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYCORE_H__
 #define __ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYCORE_H__
 
-#include "arm_compute/core/CL/kernels/CLGEMMInterleave4x4Kernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.h"
 #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/CLGEMMReshapeLHSMatrixKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h"
 #include "arm_compute/runtime/CL/CLMemoryGroup.h"
 #include "arm_compute/runtime/CL/CLTensor.h"
@@ -41,7 +41,7 @@ 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 CLGEMMReshapeLHSMatrixKernel  (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)
@@ -101,7 +101,7 @@ public:
 private:
     CLMemoryGroup                                 _memory_group;
     CLGEMMLowpMatrixMultiplyKernel                _mm_kernel;
-    CLGEMMInterleave4x4Kernel                     _mtx_a_reshape_kernel;
+    CLGEMMReshapeLHSMatrixKernel                  _mtx_a_reshape_kernel;
     CLGEMMReshapeRHSMatrixKernel                  _mtx_b_reshape_kernel;
     CLGEMMLowpMatrixAReductionKernel              _mtx_a_reduction_kernel;
     CLGEMMLowpMatrixBReductionKernel              _mtx_b_reduction_kernel;