COMPMID-1772: Implement PadV2 for NEON

Change-Id: Ia4604524a034c46b004fd850183480c5fbfd8cb3
Reviewed-on: https://review.mlplatform.org/437
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>

6 files changed, 168 insertions, 12 deletions

diff --git a/arm_compute/core/NEON/NEKernels.h b/arm_compute/core/NEON/NEKernels.h
index a99b7d08e1..b7810cbd51 100644
--- a/arm_compute/core/NEON/NEKernels.h
+++ b/arm_compute/core/NEON/NEKernels.h
@@ -94,6 +94,7 @@
 #include "arm_compute/core/NEON/kernels/NEMagnitudePhaseKernel.h"
 #include "arm_compute/core/NEON/kernels/NEMeanStdDevKernel.h"
 #include "arm_compute/core/NEON/kernels/NEMedian3x3Kernel.h"
+#include "arm_compute/core/NEON/kernels/NEMemsetKernel.h"
 #include "arm_compute/core/NEON/kernels/NEMinMaxLayerKernel.h"
 #include "arm_compute/core/NEON/kernels/NEMinMaxLocationKernel.h"
 #include "arm_compute/core/NEON/kernels/NENonLinearFilterKernel.h"
diff --git a/arm_compute/core/NEON/kernels/NEMemsetKernel.h b/arm_compute/core/NEON/kernels/NEMemsetKernel.h
new file mode 100644
index 0000000000..721ecf9ae2
--- /dev/null
+++ b/arm_compute/core/NEON/kernels/NEMemsetKernel.h
@@ -0,0 +1,71 @@
+/*
+ * Copyright (c) 2018-2019 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_NEMEMSETKERNEL_H__
+#define __ARM_COMPUTE_NEMEMSETKERNEL_H__
+
+#include "arm_compute/core/NEON/INEKernel.h"
+#include "arm_compute/core/PixelValue.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+// Forward declarations
+class ITensor;
+
+/** Interface for filling the planes of a tensor */
+class NEMemsetKernel : public INEKernel
+{
+public:
+    const char *name() const override
+    {
+        return "NEMemsetKernel";
+    }
+    /** Default constructor */
+    NEMemsetKernel();
+    /** Default destructor */
+    ~NEMemsetKernel() = default;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    NEMemsetKernel(const NEMemsetKernel &) = delete;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    NEMemsetKernel &operator=(const NEMemsetKernel &) = delete;
+    /** Allow instances of this class to be moved */
+    NEMemsetKernel(NEMemsetKernel &&) = default;
+    /** Allow instances of this class to be moved */
+    NEMemsetKernel &operator=(NEMemsetKernel &&) = default;
+    /** Initialise the kernel's tensor and filling value
+     *
+     * @param[in,out] tensor         Input tensor to fill. Supported data types: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+     * @param[in]     constant_value The value used to fill the planes of the tensor
+     */
+    void configure(ITensor *tensor, const PixelValue &constant_value);
+
+    // Inherited methods overridden:
+    void run(const Window &window, const ThreadInfo &info) override;
+
+private:
+    ITensor   *_tensor;
+    PixelValue _constant_value;
+};
+} // namespace arm_compute
+#endif /*__ARM_COMPUTE_NEMEMSETKERNEL_H__ */
diff --git a/arm_compute/core/utils/misc/ShapeCalculator.h b/arm_compute/core/utils/misc/ShapeCalculator.h
index 619234d306..f8a6df7fe7 100644
--- a/arm_compute/core/utils/misc/ShapeCalculator.h
+++ b/arm_compute/core/utils/misc/ShapeCalculator.h
@@ -1011,7 +1011,9 @@ inline TensorShape compute_padded_shape(const TensorShape &input_shape, const Pa
     TensorShape padded_shape = input_shape;
     for(size_t dim = 0; dim < padding.size(); ++dim)
     {
-        padded_shape.set(dim, padding[dim].first + input_shape[dim] + padding[dim].second);
+        const auto    &padding_pair   = padding[dim];
+        const uint32_t shape_on_index = (padded_shape.num_dimensions() <= dim) ? 1 : input_shape[dim];
+        padded_shape.set(dim, padding_pair.first + shape_on_index + padding_pair.second);
     }
     return padded_shape;
 }
diff --git a/arm_compute/runtime/CL/functions/CLPadLayer.h b/arm_compute/runtime/CL/functions/CLPadLayer.h
index b9dca665d0..1ecf82fa7c 100644
--- a/arm_compute/runtime/CL/functions/CLPadLayer.h
+++ b/arm_compute/runtime/CL/functions/CLPadLayer.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018 ARM Limited.
+ * Copyright (c) 2018-2019 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -49,21 +49,23 @@ public:
 
     /** Initialize the function
      *
-     * @param[in]  input   Source tensor. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
-     * @param[out] output  Output tensor. Data type supported: same as @p input
-     * @param[in]  padding The padding for each spatial dimension of the input tensor. The pair padding[i]
-     *                     specifies the front and the end padding in the i-th dimension.
+     * @param[in]  input          Source tensor. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
+     * @param[out] output         Output tensor. Data type supported: same as @p input
+     * @param[in]  padding        The padding for each spatial dimension of the input tensor. The pair padding[i]
+     *                            specifies the front and the end padding in the i-th dimension.
+     * @param[in]  constant_value (Optional) Constant value to be used for the padding
      */
-    void configure(ICLTensor *input, ICLTensor *output, const PaddingList &padding);
+    void configure(ICLTensor *input, ICLTensor *output, const PaddingList &padding, PixelValue constant_value = PixelValue());
 
     /**  Static function to check if given info will lead to a valid configuration of @ref CLPadLayer.
      *
-     * @param[in] input   Source tensor info. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
-     * @param[in] output  Output tensor info. Data type supported: same as @p input
-     * @param[in] padding The padding for each spatial dimension of the input tensor. The pair padding[i]
-     *                     specifies the front and the end padding in the i-th dimension.
+     * @param[in] input          Source tensor info. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
+     * @param[in] output         Output tensor info. Data type supported: same as @p input
+     * @param[in] padding        The padding for each spatial dimension of the input tensor. The pair padding[i]
+     *                           specifies the front and the end padding in the i-th dimension.
+     * @param[in] constant_value (Optional) Constant value to be used for the padding
      */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PaddingList &padding);
+    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PaddingList &padding, PixelValue constant_value = PixelValue());
 
     // Inherited methods overridden:
     void run() override;
diff --git a/arm_compute/runtime/NEON/NEFunctions.h b/arm_compute/runtime/NEON/NEFunctions.h
index ceebc7b844..606f282e16 100644
--- a/arm_compute/runtime/NEON/NEFunctions.h
+++ b/arm_compute/runtime/NEON/NEFunctions.h
@@ -99,6 +99,7 @@
 #include "arm_compute/runtime/NEON/functions/NENonMaximaSuppression3x3.h"
 #include "arm_compute/runtime/NEON/functions/NENormalizationLayer.h"
 #include "arm_compute/runtime/NEON/functions/NEOpticalFlow.h"
+#include "arm_compute/runtime/NEON/functions/NEPadLayer.h"
 #include "arm_compute/runtime/NEON/functions/NEPermute.h"
 #include "arm_compute/runtime/NEON/functions/NEPhase.h"
 #include "arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h"
diff --git a/arm_compute/runtime/NEON/functions/NEPadLayer.h b/arm_compute/runtime/NEON/functions/NEPadLayer.h
new file mode 100644
index 0000000000..3a0863802a
--- /dev/null
+++ b/arm_compute/runtime/NEON/functions/NEPadLayer.h
@@ -0,0 +1,79 @@
+/*
+ * Copyright (c) 2018-2019 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_NEPADLAYER_H__
+#define __ARM_COMPUTE_NEPADLAYER_H__
+
+#include "arm_compute/runtime/IFunction.h"
+#include "arm_compute/runtime/SubTensor.h"
+
+#include "arm_compute/core/NEON/kernels/NECopyKernel.h"
+#include "arm_compute/core/NEON/kernels/NEMemsetKernel.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+// Forward declarations
+class ITensor;
+
+/** Basic function to pad a tensor. This function calls the following NEON kernels:
+ *
+ *  -# @ref NEMemsetKernel
+ *  -# @ref NECopyKernel
+ */
+class NEPadLayer : public IFunction
+{
+public:
+    /** Default constructor*/
+    NEPadLayer();
+    /** Initialize the function
+     *
+     * @param[in]  input          Source tensor. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
+     * @param[out] output         Output tensor. Data type supported: same as @p input
+     * @param[in]  padding        The padding for each spatial dimension of the input tensor. The pair padding[i]
+     *                            specifies the front and the end padding in the i-th dimension.
+     * @param[in]  constant_value (Optional) Constant value to be used for the padding
+     */
+    void configure(ITensor *input, ITensor *output, const PaddingList &padding, PixelValue constant_value = PixelValue());
+    /**  Static function to check if given info will lead to a valid configuration of @ref NEPadLayer.
+     *
+     * @param[in] input          Source tensor info. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
+     * @param[in] output         Output tensor info. Data type supported: same as @p input
+     * @param[in] padding        The padding for each spatial dimension of the input tensor. The pair padding[i]
+     *                           specifies the front and the end padding in the i-th dimension.
+     * @param[in] constant_value (Optional) Constant value to be used for the padding
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PaddingList &padding, PixelValue constant_value = PixelValue());
+
+    // Inherited methods overridden:
+    void run() override;
+
+private:
+    NEMemsetKernel _memset_kernel;
+    NECopyKernel   _copy_kernel;
+    SubTensor      _output_subtensor;
+};
+} // namespace arm_compute
+#endif /*__ARM_COMPUTE_NEPADLAYER_H__ */