COMPMID-1932: Implement CLHeightConcatenateLayer.

Change-Id: I244ff2352c3c9620148dab6475aee62b218b96ef
Signed-off-by: Pablo Tello <pablo.tello@arm.com>
Reviewed-on: https://review.mlplatform.org/c/871
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>

4 files changed, 97 insertions, 6 deletions

diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
index 396158ff2d..2fd2341e48 100644
--- a/arm_compute/core/CL/CLKernels.h
+++ b/arm_compute/core/CL/CLKernels.h
@@ -95,6 +95,7 @@
 #include "arm_compute/core/CL/kernels/CLHOGDescriptorKernel.h"
 #include "arm_compute/core/CL/kernels/CLHOGDetectorKernel.h"
 #include "arm_compute/core/CL/kernels/CLHarrisCornersKernel.h"
+#include "arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h"
 #include "arm_compute/core/CL/kernels/CLHistogramKernel.h"
 #include "arm_compute/core/CL/kernels/CLIm2ColKernel.h"
 #include "arm_compute/core/CL/kernels/CLIntegralImageKernel.h"
diff --git a/arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h b/arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h
new file mode 100644
index 0000000000..a9b2f036bc
--- /dev/null
+++ b/arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h
@@ -0,0 +1,81 @@
+/*
+ * 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_CLHEIGHTCONCATENATELAYERKERNEL_H__
+#define __ARM_COMPUTE_CLHEIGHTCONCATENATELAYERKERNEL_H__
+
+#include "arm_compute/core/CL/ICLKernel.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+class ICLTensor;
+
+/** Interface for the height concatenate kernel.
+ *  The input tensor will be concatenated into the output tensor.
+ */
+class CLHeightConcatenateLayerKernel : public ICLKernel
+{
+public:
+    /** Default constructor */
+    CLHeightConcatenateLayerKernel();
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLHeightConcatenateLayerKernel(const CLHeightConcatenateLayerKernel &) = delete;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLHeightConcatenateLayerKernel &operator=(const CLHeightConcatenateLayerKernel &) = delete;
+    /** Allow instances of this class to be moved */
+    CLHeightConcatenateLayerKernel(CLHeightConcatenateLayerKernel &&) = default;
+    /** Allow instances of this class to be moved */
+    CLHeightConcatenateLayerKernel &operator=(CLHeightConcatenateLayerKernel &&) = default;
+    /** Default destructor */
+    ~CLHeightConcatenateLayerKernel() = default;
+    /** Initialise the kernel's inputs and output
+     *
+     * @param[in]  input         Input tensor. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+     * @param[in]  height_offset The starting offset on the Y axis for the output tensor.
+     * @param[out] output        Output tensor. Data types supported: Same as @p input.
+     *
+     */
+    void configure(const ICLTensor *input, unsigned int height_offset, ICLTensor *output);
+    /**  Static function to check if given info will lead to a valid configuration of @ref CLHeightConcatenateLayerKernel
+     *
+     * @param[in] input         Input tensor info. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+     * @param[in] height_offset The starting offset on the Y axis for the output tensor.
+     * @param[in] output        Output tensor info. Data types supported: Same as @p input.
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *input, unsigned int height_offset, const ITensorInfo *output);
+
+    // Inherited methods overridden:
+    void run(const Window &window, cl::CommandQueue &queue) override;
+
+private:
+    const ICLTensor *_input;
+    ICLTensor       *_output;
+    unsigned int     _height_offset;
+    unsigned int     _num_elems_processed_per_iteration;
+};
+} // namespace arm_compute
+#endif /* __ARM_COMPUTE_CLHEIGHTCONCATENATELAYERKERNEL_H__ */
diff --git a/arm_compute/core/utils/misc/ShapeCalculator.h b/arm_compute/core/utils/misc/ShapeCalculator.h
index 6782cda7fe..7a34b43028 100644
--- a/arm_compute/core/utils/misc/ShapeCalculator.h
+++ b/arm_compute/core/utils/misc/ShapeCalculator.h
@@ -1239,6 +1239,7 @@ inline TensorShape calculate_concatenate_shape(const std::vector<T *> &input, si
     {
         ARM_COMPUTE_ERROR_ON(tensor == nullptr);
         const TensorShape shape = extract_shape(tensor);
+        ARM_COMPUTE_ERROR_ON(axis >= shape.num_dimensions());
         new_size += shape[axis];
     }
 
diff --git a/arm_compute/runtime/CL/functions/CLConcatenateLayer.h b/arm_compute/runtime/CL/functions/CLConcatenateLayer.h
index 4d4c62434a..25a0153f51 100644
--- a/arm_compute/runtime/CL/functions/CLConcatenateLayer.h
+++ b/arm_compute/runtime/CL/functions/CLConcatenateLayer.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018 ARM Limited.
+ * Copyright (c) 2018-2019 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -26,6 +26,7 @@
 
 #include "arm_compute/runtime/IFunction.h"
 
+#include "arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h"
 #include "arm_compute/core/Types.h"
 
 #include <memory>
@@ -41,6 +42,7 @@ class Status;
 /** Basic function to execute concatenate tensors along a given axis. This function calls the following kernels:
  *
  * -# @ref CLWidthConcatenateLayer (if underlying concatenation axis is 0).
+ * -# @ref CLHeightConcatenateLayerKernel (if underlying concatenation axis is 1).
  * -# @ref CLDepthConcatenateLayer (if underlying concatenation axis is 2).
  */
 class CLConcatenateLayer : public IFunction
@@ -51,21 +53,21 @@ public:
     /** Initialise the kernel's inputs vector and output.
      *
      * @note Input and output tensor dimensions preconditions defer depending on the concatenation axis.
-     * @note Preconditions can be found respectively at @ref CLWidthConcatenateLayer and @ref CLDepthConcatenateLayer.
+     * @note Preconditions can be found respectively at @ref CLWidthConcatenateLayer, @ref CLHeightConcatenateLayerKernel and @ref CLDepthConcatenateLayer.
      *
      * @param[in,out] inputs_vector The vectors containing all the tensors to concatenate. Data types supported: QASYMM8/F16/F32.
      * @param[out]    output        Output tensor. Data types supported: Same as @p input.
-     * @param[in]     axis          Concatenation axis. Supported underlying concatenation axis are 0 and 2.
+     * @param[in]     axis          Concatenation axis. Supported underlying concatenation axis are 0, 1 and 2.
      */
     void configure(const std::vector<ICLTensor *> &inputs_vector, ICLTensor *output, DataLayoutDimension axis);
     /** Static function to check if given info will lead to a valid configuration of @ref CLConcatenateLayer
      *
      * @note Input and output tensor dimensions preconditions defer depending on the concatenation axis.
-     * @note Preconditions can be found respectively at @ref CLWidthConcatenateLayer and @ref CLDepthConcatenateLayer.
+     * @note Preconditions can be found respectively at @ref CLWidthConcatenateLayer, @ref CLHeightConcatenateLayerKernel and @ref CLDepthConcatenateLayer.
      *
      * @param[in] inputs_vector The vectors containing all the tensors info to concatenate. Data types supported: QASYMM8/F16/F32.
      * @param[in] output        Output tensor info. Data types supported: Same as @p input.
-     * @param[in] axis          Concatenation axis. Supported underlying concatenation axis are 0 and 2.
+     * @param[in] axis          Concatenation axis. Supported underlying concatenation axis are 0, 1 and 2.
      *
      * @return a status
      */
@@ -75,7 +77,13 @@ public:
     void run() override;
 
 private:
-    std::unique_ptr<IFunction> _concat_function;
+    void configure_h_concatenate(std::vector<ICLTensor *> inputs_vector, ICLTensor *output);
+    static Status validate_h_concatenate(const std::vector<ITensorInfo *> &inputs_vector, const ITensorInfo *output);
+
+    std::unique_ptr<IFunction>                        _concat_function;
+    std::unique_ptr<CLHeightConcatenateLayerKernel[]> _hconcat_kernels;
+    unsigned int                                      _num_inputs;
+    unsigned int                                      _axis;
 };
 }
 #endif /* __ARM_COMPUTE_CLCONCATENATELAYER_H__ */