COMPMID-2002: Implement CLGEMMLowpMatrixMultiplyReshapedOnlyRHS - Transposed

Change-Id: I3907d151107766dc34749fe5710d7219e810b39f
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/875
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Giuseppe Rossini <giuseppe.rossini@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>

3 files changed, 98 insertions, 1 deletions

diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
index b767812fc8..e3ffcd0704 100644
--- a/arm_compute/core/CL/CLKernels.h
+++ b/arm_compute/core/CL/CLKernels.h
@@ -74,6 +74,7 @@
 #include "arm_compute/core/CL/kernels/CLGEMMInterleave4x4Kernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h"
+#include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.h"
 #include "arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h"
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h
index 139f7ab0a8..eaadaeff19 100644
--- a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h
+++ b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h
@@ -100,4 +100,4 @@ private:
     bool             _use_dummy_work_items;
 };
 } // namespace arm_compute
-#endif /*__ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDKERNEL_H__*/
\ No newline at end of file
+#endif /*__ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDKERNEL_H__*/
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h
new file mode 100644
index 0000000000..1fd987528c
--- /dev/null
+++ b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h
@@ -0,0 +1,96 @@
+/*
+ * Copyright (c) 2019 ARM Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef __ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H__
+#define __ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H__
+
+#include "arm_compute/core/CL/ICLKernel.h"
+
+namespace arm_compute
+{
+class ICLTensor;
+
+/** OpenCL kernel to multiply matrices with QASYMM8 data type when only the input matrix RHS (input1) has been reshaped
+ *
+ * @note The input matrix input1 must be reshaped through @ref CLGEMMReshapeRHSMatrixKernel
+ */
+class CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel : public ICLKernel
+{
+public:
+    /** Default Constructor */
+    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel();
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel(const CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &) = delete;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &operator=(const CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &) = delete;
+    /** Allow instances of this class to be moved */
+    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel(CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &&) = default;
+    /** Allow instances of this class to be moved */
+    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &operator=(CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &&) = default;
+    /** Initialise the kernel's input and output.
+     *
+     * @param[in]  input0    Input tensor containing the LHS matrix. Data type supported: QASYMM8
+     * @param[in]  input1    Input tensor containing the RHS reshaped matrix. 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]  lhs_info  LHS matrix information used to retrieve the number of rows to be processed by each thread
+     *                       lhs_info.m0: 2,3,4,5,6,7,8
+     *                       lhs_info.k0: 2,3,4,8,16
+     * @param[in]  rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
+     *                       rhs_info.n0: 2,3,4,8,16
+     *                       rhs_info.k0: 2,3,4,8,16
+     *                       rhs_info.transpose: true
+     * @param[in]  gemm_info GEMM information used to retrieve the original dimensions of the input matrices
+     */
+    void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info, const GEMMReshapeInfo &gemm_info);
+    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel
+     *
+     * @param[in] input0    Input tensor info for the LHS matrix. Data type supported: QASYMM8
+     * @param[in] input1    Input tensor info for the RHS reshaped matrix. Data type supported: same as @p input0
+     * @param[in] output    Output tensor info. Data type supported: S32
+     * @param[in] lhs_info  LHS matrix information used to retrieve the number of rows to be processed by each thread
+     *                      lhs_info.m0: 2,3,4,5,6,7,8
+     *                      lhs_info.k0: 2,3,4,8,16
+     * @param[in] rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
+     *                      rhs_info.n0: 2,3,4,8,16
+     *                      rhs_info.k0: 2,3,4,8,16
+     *                      rhs_info.transpose: true
+     * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info,
+                           const GEMMReshapeInfo &gemm_info);
+
+    // Inherited methods overridden:
+    void run(const Window &window, cl::CommandQueue &queue) override;
+
+private:
+    const ICLTensor *_input0;
+    const ICLTensor *_input1;
+    ICLTensor       *_output;
+    bool             _slide_matrix_b;
+    bool             _reinterpret_input_as_3d;
+    bool             _reinterpret_output_as_3d;
+};
+} // namespace arm_compute
+#endif /*__ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H__*/
\ No newline at end of file