COMPMID-1137 OpenCL concatenate width
Change-Id: I40faba421281b1cf080fa6a825d04a4366cdaeb0
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/130700
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
index 9316c6e..f1171a8 100644
--- a/arm_compute/core/CL/CLKernels.h
+++ b/arm_compute/core/CL/CLKernels.h
@@ -113,6 +113,7 @@
#include "arm_compute/core/CL/kernels/CLWarpAffineKernel.h"
#include "arm_compute/core/CL/kernels/CLWarpPerspectiveKernel.h"
#include "arm_compute/core/CL/kernels/CLWeightsReshapeKernel.h"
+#include "arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h"
#include "arm_compute/core/CL/kernels/CLWinogradFilterTransformKernel.h"
#include "arm_compute/core/CL/kernels/CLWinogradInputTransformKernel.h"
#include "arm_compute/core/CL/kernels/CLWinogradOutputTransformKernel.h"
diff --git a/arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h b/arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h
new file mode 100644
index 0000000..5b8a318
--- /dev/null
+++ b/arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h
@@ -0,0 +1,80 @@
+/*
+ * Copyright (c) 2018 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_CLWIDTHCONCATENATELAYERKERNEL_H__
+#define __ARM_COMPUTE_CLWIDTHCONCATENATELAYERKERNEL_H__
+
+#include "arm_compute/core/CL/ICLKernel.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+class ICLTensor;
+
+/** Interface for the width concatenate kernel.
+ * The input tensor will be concatenated into the output tensor.
+ */
+class CLWidthConcatenateLayerKernel : public ICLKernel
+{
+public:
+ /** Default constructor */
+ CLWidthConcatenateLayerKernel();
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ CLWidthConcatenateLayerKernel(const CLWidthConcatenateLayerKernel &) = delete;
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ CLWidthConcatenateLayerKernel &operator=(const CLWidthConcatenateLayerKernel &) = delete;
+ /** Allow instances of this class to be moved */
+ CLWidthConcatenateLayerKernel(CLWidthConcatenateLayerKernel &&) = default;
+ /** Allow instances of this class to be moved */
+ CLWidthConcatenateLayerKernel &operator=(CLWidthConcatenateLayerKernel &&) = default;
+ /** Default destructor */
+ ~CLWidthConcatenateLayerKernel() = default;
+ /** Initialise the kernel's inputs and output
+ *
+ * @param[in] input Input tensor. Data types supported: U8/S8/QS8/QASYMM8/U16/S16/QS16/F16/U32/S32/F32
+ * @param[in] width_offset The offset on the X axis.
+ * @param[in,out] output Output tensor. Data types supported: Same as @p input.
+ *
+ */
+ void configure(const ICLTensor *input, unsigned int width_offset, ICLTensor *output);
+ /** Static function to check if given info will lead to a valid configuration of @ref CLDepthConcatenateLayerKernel
+ *
+ * @param[in] input Input tensor info. Data types supported: U8/S8/QS8/QASYMM8/U16/S16/QS16/F16/U32/S32/F32
+ * @param[in] width_offset The offset on the X axis.
+ * @param[in] output Output tensor info. Data types supported: Same as @p input.
+ *
+ * @return a status
+ */
+ static Status validate(const ITensorInfo *input, unsigned int width_offset, const ITensorInfo *output);
+
+ // Inherited methods overridden:
+ void run(const Window &window, cl::CommandQueue &queue) override;
+
+private:
+ const ICLTensor *_input;
+ ICLTensor *_output;
+ unsigned int _width_offset;
+};
+} // namespace arm_compute
+#endif /* __ARM_COMPUTE_CLWIDTHCONCATENATELAYERKERNEL_H__ */
diff --git a/arm_compute/core/utils/misc/ShapeCalculator.h b/arm_compute/core/utils/misc/ShapeCalculator.h
index 30d3f9b..deab181 100644
--- a/arm_compute/core/utils/misc/ShapeCalculator.h
+++ b/arm_compute/core/utils/misc/ShapeCalculator.h
@@ -366,6 +366,35 @@
return tensor_shape;
}
+
+template <typename T>
+inline TensorShape get_shape_from_info(T *info)
+{
+ return info->info()->tensor_shape();
+}
+
+inline TensorShape get_shape_from_info(ITensorInfo *info)
+{
+ return info->tensor_shape();
+}
+
+template <typename T>
+inline TensorShape calculate_width_concatenate_shape(const std::vector<T *> &inputs_vector)
+{
+ TensorShape out_shape = get_shape_from_info(inputs_vector[0]);
+
+ size_t width = 0;
+ for(const auto &tensor : inputs_vector)
+ {
+ ARM_COMPUTE_ERROR_ON(tensor == nullptr);
+ const TensorShape shape = get_shape_from_info(tensor);
+ width += shape.x();
+ }
+
+ out_shape.set(0, width);
+
+ return out_shape;
+}
} // namespace shape_calculator
} // namespace misc
} // namespace arm_compute