COMPMID-355 Implement 3x3 CL direct convolution

Change-Id: I1b44dc375045964e65557f0ead57a7c12d6bf097
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/81418
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>

2 files changed, 87 insertions, 0 deletions

diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
index 0e9f356e52..cf36beffa3 100644
--- a/arm_compute/core/CL/CLKernels.h
+++ b/arm_compute/core/CL/CLKernels.h
@@ -46,6 +46,7 @@
 #include "arm_compute/core/CL/kernels/CLDepthConvertKernel.h"
 #include "arm_compute/core/CL/kernels/CLDerivativeKernel.h"
 #include "arm_compute/core/CL/kernels/CLDilateKernel.h"
+#include "arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h"
 #include "arm_compute/core/CL/kernels/CLErodeKernel.h"
 #include "arm_compute/core/CL/kernels/CLFastCornersKernel.h"
 #include "arm_compute/core/CL/kernels/CLFillBorderKernel.h"
diff --git a/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h b/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
new file mode 100644
index 0000000000..28eecf029a
--- /dev/null
+++ b/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
@@ -0,0 +1,86 @@
+/*
+ * Copyright (c) 2017 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_CLDIRECTCONVOLUTIONLAYERKERNEL_H__
+#define __ARM_COMPUTE_CLDIRECTCONVOLUTIONLAYERKERNEL_H__
+
+#include "arm_compute/core/CL/ICLKernel.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+class ICLTensor;
+
+/** Interface for the  direct convolution kernel.
+ */
+template <unsigned int kernel_size>
+class CLDirectConvolutionLayerKernel : public ICLKernel
+{
+public:
+    /** Default constructor */
+    CLDirectConvolutionLayerKernel();
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLDirectConvolutionLayerKernel(const CLDirectConvolutionLayerKernel &) = delete;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLDirectConvolutionLayerKernel &operator=(const CLDirectConvolutionLayerKernel &) = delete;
+    /** Allow instances of this class to be moved */
+    CLDirectConvolutionLayerKernel(CLDirectConvolutionLayerKernel &&) = default;
+    /** Allow instances of this class to be moved */
+    CLDirectConvolutionLayerKernel &operator=(CLDirectConvolutionLayerKernel &&) = default;
+    /** Default destructor */
+    ~CLDirectConvolutionLayerKernel() = default;
+    /** Set the input, weights, biases and output tensors.
+     *
+     * @param[in]  input     The input tensor to convolve. 3 lower dimensions represent a single input [width, height, IFM],
+     *                       while every optional dimension from 4 and above represent a batch of inputs. Data types supported: F32.
+     * @param[in]  weights   Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM].
+     *                       The 3rd dimension must be the same as the input's volume 3rd dimension.
+     *                       Data type supported:Same as @p input.
+     * @param[in]  biases    Biases tensor. Biases are 1D tensor with dimension [OFM]. Data type supported: Same as @p input.
+     * @param[out] output    Output tensor.
+     *                       The 3rd dimensions must be equal to the 4th dimension of the @p kernels tensor. Data types supported: Same as @p input.
+     * @param[in]  conv_info Contains padding and stride information described in @ref PadStrideInfo.
+     */
+    void configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info);
+
+    // Inherited methods overridden:
+    BorderSize border_size() const override;
+
+    // Inherited methods overridden:
+    void run(const Window &window, cl::CommandQueue &queue) override;
+
+private:
+    const ICLTensor *_input;
+    const ICLTensor *_biases;
+    const ICLTensor *_weights;
+    ICLTensor       *_output;
+    BorderSize       _border_size;
+    int              _conv_pad_x;
+    int              _conv_pad_y;
+    int              _conv_stride_x;
+    int              _conv_stride_y;
+};
+
+using CLDirectConvolutionLayer3x3Kernel = CLDirectConvolutionLayerKernel<3>;
+}
+#endif /*__ARM_COMPUTE_CLDIRECTCONVOLUTIONLAYERKERNEL_H__ */