COMPMID-791: Generic Depthwise Convolution Layer NEON QASYMM8

Change-Id: I33cf54e68f6c097ac58b6f16c3f9a720978f09cd
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/117289
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>

1 files changed, 20 insertions, 6 deletions

diff --git a/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h b/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h
index 8d59ba3248..ca10bfaab2 100644
--- a/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h
+++ b/arm_compute/core/NEON/kernels/NEDepthwiseIm2ColKernel.h
@@ -55,7 +55,7 @@ public:
     /** Set the input and output of the kernel.
      *
      * @param[in]  input       The input tensor to convert. 3 lower dimensions represent a single input [width, height, IFM],
-     *                         while every optional dimension from 4 and above represent a batch of inputs. Data types supported: F32
+     *                         while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8, F32
      * @param[out] output      The output tensor. First 3 lower dimensions represent a transform of each 3D input,
      *                         while every dimension above 3 represents a batch. Data types supported: Same as @p input
      * @param[in]  kernel_dims The kernel dimensions (width and height).
@@ -68,11 +68,25 @@ public:
     void run(const Window &window, const ThreadInfo &info) override;
 
 private:
-    const ITensor *_input;
-    ITensor       *_output;
-    Size2D         _kernel_dims;
-    PadStrideInfo  _conv_info;
-    bool           _has_bias;
+    /** Template function to run the im2col used for the depthwise convolution layer case
+     *
+     * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
+     */
+    template <typename T>
+    void run_generic(const Window &window);
+    /** Common signature for all the specialised depthwise im2col functions
+     *
+     * @param[in] window Region on which to execute the kernel.
+     */
+    using DepthwiseIm2ColFunctionPtr = void (NEDepthwiseIm2ColKernel::*)(const Window &window);
+
+private:
+    DepthwiseIm2ColFunctionPtr _func;
+    const ITensor             *_input;
+    ITensor                   *_output;
+    Size2D                     _kernel_dims;
+    PadStrideInfo              _conv_info;
+    bool                       _has_bias;
 };
 } // arm_compute
 #endif /*__ARM_COMPUTE_NEDEPTHWISEIM2COLKERNEL_H__ */