COMPMID-1720: CL: Implement Tile
Change-Id: I2a18f0acea382960a8bc71a8f56928a5998f0dd6
diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
index 1457d7b..a045322 100644
--- a/arm_compute/core/CL/CLKernels.h
+++ b/arm_compute/core/CL/CLKernels.h
@@ -125,6 +125,7 @@
#include "arm_compute/core/CL/kernels/CLStridedSliceKernel.h"
#include "arm_compute/core/CL/kernels/CLTableLookupKernel.h"
#include "arm_compute/core/CL/kernels/CLThresholdKernel.h"
+#include "arm_compute/core/CL/kernels/CLTileKernel.h"
#include "arm_compute/core/CL/kernels/CLTransposeKernel.h"
#include "arm_compute/core/CL/kernels/CLUpsampleLayerKernel.h"
#include "arm_compute/core/CL/kernels/CLWarpAffineKernel.h"
diff --git a/arm_compute/core/CL/kernels/CLTileKernel.h b/arm_compute/core/CL/kernels/CLTileKernel.h
new file mode 100644
index 0000000..e575b7b
--- /dev/null
+++ b/arm_compute/core/CL/kernels/CLTileKernel.h
@@ -0,0 +1,78 @@
+/*
+ * 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_CLTILEKERNEL_H__
+#define __ARM_COMPUTE_CLTILEKERNEL_H__
+
+#include "arm_compute/core/CL/ICLKernel.h"
+
+namespace arm_compute
+{
+class ICLTensor;
+
+/** OpenCL kernel to perform a Tile operation */
+class CLTileKernel : public ICLKernel
+{
+public:
+ /** Default constructor */
+ CLTileKernel();
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ CLTileKernel(const CLTileKernel &) = delete;
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ CLTileKernel &operator=(const CLTileKernel &) = delete;
+ /** Allow instances of this class to be moved */
+ CLTileKernel(CLTileKernel &&) = default;
+ /** Allow instances of this class to be moved */
+ CLTileKernel &operator=(CLTileKernel &&) = default;
+ /** Default destructor */
+ ~CLTileKernel() = default;
+ /** Set the source, destination of the kernel
+ *
+ * @param[in] input Source tensor. Data type supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
+ * @param[in] multiples Contains the number of times the input tensor should be replicated on the given dimension.
+ * Cannot have more than 4 elements (tiling in dimensions greater than 4 is not supported).
+ * @param[out] output Destination tensor. Same as @p input
+ *
+ */
+ void configure(const ICLTensor *input, ICLTensor *output, const Multiples &multiples);
+ /** Static function to check if given info will lead to a valid configuration of @ref CLTileKernel
+ *
+ * @param[in] input Source tensor info. Data type supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32.
+ * @param[in] multiples Contains the number of times the input tensor should be replicated on the given dimension.
+ * Cannot have more than 4 elements (tiling in dimensions greater than 4 is not supported).
+ * @param[in] output Destination tensor info. Same as @p input
+ *
+ *
+ * @return a status
+ */
+ static Status validate(const ITensorInfo *input, const ITensorInfo *output, const Multiples &multiples);
+
+ // Inherited methods overridden:
+ void run(const Window &window, cl::CommandQueue &queue) override;
+
+private:
+ const ICLTensor *_input;
+ ICLTensor *_output;
+};
+} // namespace arm_compute
+#endif /*__ARM_COMPUTE_CLTILEKERNEL_H__ */
diff --git a/arm_compute/core/Types.h b/arm_compute/core/Types.h
index 9f3857c..d6122a6 100644
--- a/arm_compute/core/Types.h
+++ b/arm_compute/core/Types.h
@@ -479,6 +479,9 @@
/** List of padding information */
using PaddingList = std::vector<PaddingInfo>;
+/** Information to produce a tiled version of a Tensor */
+using Multiples = std::vector<uint32_t>;
+
/** Region of interest */
struct ROI
{
diff --git a/arm_compute/core/utils/misc/ShapeCalculator.h b/arm_compute/core/utils/misc/ShapeCalculator.h
index da9ff56..38906df 100644
--- a/arm_compute/core/utils/misc/ShapeCalculator.h
+++ b/arm_compute/core/utils/misc/ShapeCalculator.h
@@ -671,6 +671,16 @@
return padded_shape;
}
+inline TensorShape compute_tiled_shape(const TensorShape &input_shape, const Multiples &multiples)
+{
+ TensorShape tiled_shape = input_shape;
+ for(size_t dim = 0; dim < multiples.size(); ++dim)
+ {
+ tiled_shape.set(dim, input_shape[dim] * multiples[dim]);
+ }
+ return tiled_shape;
+}
+
inline TensorShape compute_upsample_shape(const ITensorInfo &input, const Size2D &info)
{
const DataLayout data_layout = input.data_layout();