Implement the OpenCL kernel to compute the indirect convolution

- Implement indirect convolution kernel
- Add operator support
- Add test

Resolves COMPMID-5709

Change-Id: I9272304163471a5a40da7fdec204599f3c1d8e32
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8701
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

2 files changed, 121 insertions, 0 deletions

diff --git a/arm_compute/runtime/CL/CLFunctions.h b/arm_compute/runtime/CL/CLFunctions.h
index f42da5801c..e37134d454 100644
--- a/arm_compute/runtime/CL/CLFunctions.h
+++ b/arm_compute/runtime/CL/CLFunctions.h
@@ -69,6 +69,7 @@
 #include "arm_compute/runtime/CL/functions/CLGEMMLowpOutputStage.h"
 #include "arm_compute/runtime/CL/functions/CLGather.h"
 #include "arm_compute/runtime/CL/functions/CLGenerateProposalsLayer.h"
+#include "arm_compute/runtime/CL/functions/CLIndirectConvolutionLayer.h"
 #include "arm_compute/runtime/CL/functions/CLInstanceNormalizationLayer.h"
 #include "arm_compute/runtime/CL/functions/CLL2NormalizeLayer.h"
 #include "arm_compute/runtime/CL/functions/CLLSTMLayer.h"
diff --git a/arm_compute/runtime/CL/functions/CLIndirectConvolutionLayer.h b/arm_compute/runtime/CL/functions/CLIndirectConvolutionLayer.h
new file mode 100644
index 0000000000..8185f8df78
--- /dev/null
+++ b/arm_compute/runtime/CL/functions/CLIndirectConvolutionLayer.h
@@ -0,0 +1,120 @@
+/*
+ * Copyright (c) 2022 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_CLINDIRECTCONVOLUTIONLAYER_H
+#define ARM_COMPUTE_CLINDIRECTCONVOLUTIONLAYER_H
+
+#include "arm_compute/core/Types.h"
+#include "arm_compute/runtime/IFunction.h"
+
+#include <memory>
+
+namespace arm_compute
+{
+class CLCompileContext;
+class ICLTensor;
+class ITensorInfo;
+
+/** Basic function to run the indirect convolution function
+ */
+class CLIndirectConvolutionLayer : public IFunction
+{
+public:
+    /** Constructor */
+    CLIndirectConvolutionLayer();
+    /** Destructor */
+    ~CLIndirectConvolutionLayer();
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLIndirectConvolutionLayer(const CLIndirectConvolutionLayer &) = delete;
+    /** Default move constructor */
+    CLIndirectConvolutionLayer(CLIndirectConvolutionLayer &&);
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    CLIndirectConvolutionLayer &operator=(const CLIndirectConvolutionLayer &) = delete;
+    /** Default move assignment operator */
+    CLIndirectConvolutionLayer &operator=(CLIndirectConvolutionLayer &&);
+    /** Set the input and output tensors.
+     *
+     * Valid data layouts:
+     * - NHWC
+     *
+     * Valid data type configurations:
+     * |src0           |src1           |src2   |dst            |
+     * |:--------------|:--------------|:------|:--------------|
+     * |F16            |F16            |F16    |F16            |
+     * |F32            |F32            |F32    |F32            |
+     *
+     * @param[in]  input     Source tensor. 3 lower dimensions represent a single input,
+     *                       while every optional dimension from 4 and above represent a batch of inputs.
+     *                       Data types supported: F16/F32.
+     * @param[in]  weights   Weights tensor. Weights are 4D tensor. Data type supported:Same as @p input.
+     * @param[in]  biases    Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM].
+     *                       Data type supported: Should match @p input data type.
+     * @param[out] output    Destination tensor. 3 lower dimensions represent a single output, while the rest represent batch of outputs.
+     *                       Data types supported: Same as @p input.
+     * @param[in]  conv_info Contains padding and stride information described in @ref PadStrideInfo.
+     * @param[in]  act_info  (Optional) Activation layer information in case of a fused activation.
+     */
+    void configure(ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info, const ActivationLayerInfo &act_info = ActivationLayerInfo());
+    /** Set the input and output tensors.
+     *
+     * @param[in]  compile_context The compile context to be used.
+     * @param[in]  input           Source tensor. 3 lower dimensions represent a single input,
+     *                             while every optional dimension from 4 and above represent a batch of inputs.
+     *                             Data types supported: F16/F32.
+     * @param[in]  weights         Weights tensor. Weights are 4D tensor with dimensions. Data type supported:Same as @p input.
+     * @param[in]  biases          Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM].
+     *                             Data type supported: Should match @p input data type.
+     * @param[out] output          Destination tensor. 3 lower dimensions represent a single output, while the rest represent batch of outputs.
+     *                             Data types supported: Same as @p input.
+     * @param[in]  conv_info       Contains padding and stride information described in @ref PadStrideInfo.
+     * @param[in]  act_info        (Optional) Activation layer information in case of a fused activation.
+     */
+    void configure(const CLCompileContext &compile_context, ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
+                   const ActivationLayerInfo &act_info = ActivationLayerInfo());
+    /** Static function to check if given info will lead to a valid configuration of @ref CLIndirectConvolutionLayer
+     *
+     * @param[in] input     Source tensor. 3 lower dimensions represent a single input,
+     *                      while every optional dimension from 4 and above represent a batch of inputs.
+     *                      Data types supported: F16/F32.
+     * @param[in] weights   Weights tensor. Weights are 4D tensor with dimensions. Data type supported:Same as @p input.
+     * @param[in] biases    Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM].
+     *                      Data type supported: Should match @p input data type.
+     * @param[in] output    Destination tensor. 3 lower dimensions represent a single output, while the rest represent batch of outputs.
+     *                      Data types supported: Same as @p input.
+     * @param[in] conv_info Contains padding and stride information described in @ref PadStrideInfo.
+     * @param[in] act_info  (Optional) Activation layer information in case of a fused activation.
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, const PadStrideInfo &conv_info,
+                           const ActivationLayerInfo &act_info = ActivationLayerInfo());
+
+    // Inherited methods overridden:
+    void run() override;
+
+private:
+    struct Impl;
+    std::unique_ptr<Impl> _impl;
+};
+}
+#endif /* ARM_COMPUTE_CLINDIRECTCONVOLUTIONLAYER_H */