Implement FP GPU depthwise convolution 1x1 kernel for in-place computation

* Implement in-place graph node mutator for 1x1 depthwise convolution

* Add in-place to validation fixture except for
  DepthwiseConvolutionLayerNativeValidationFixture as it would be a
  duplicate test otherwise (DepthwiseConvolutionLayerNative test tests
  the underlying kernel)

Resolves: COMPMID-4432

Change-Id: Id7f10f5ebdce7d49f550c0b62dbaaab7f5b59d29
Signed-off-by: SiCongLi <sicong.li@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5874
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>

1 files changed, 10 insertions, 25 deletions

diff --git a/arm_compute/runtime/CL/functions/CLDepthwiseConvolutionLayer.h b/arm_compute/runtime/CL/functions/CLDepthwiseConvolutionLayer.h
index f31a17d9cb..bbb00a1ebc 100644
--- a/arm_compute/runtime/CL/functions/CLDepthwiseConvolutionLayer.h
+++ b/arm_compute/runtime/CL/functions/CLDepthwiseConvolutionLayer.h
@@ -72,48 +72,33 @@ public:
      * |QASYMM8_SIGNED |QASYMM8_SIGNED     |S32    |QASYMM8_SIGNED |
      * |QASYMM8_SIGNED |QSYMM8_PER_CHANNEL |S32    |QASYMM8_SIGNED |
      *
+     * @param[in]      compile_context  The compile context to be used.
      * @param[in, out] input            Source tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/FP16/FP32. Data layout supported: NHWC, NCHW
      * @param[in]      weights          Weights tensor. These are 3D tensors with shape [kernel_x, kernel_y, IFM].
      *                                  Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
      * @param[in]      biases           Biases tensor. A 1D tensor with shape [IFM]. Must be nullptr if not needed.
      *                                  Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out]     output           Destination tensor. Data type supported: same as @p input.
+     * @param[out]     output           Destination tensor. Pass in nullptr or @p input for in-place operation. Data type supported: same as @p input.
      * @param[in]      conv_info        Padding and stride information to use for the convolution.
      * @param[in]      depth_multiplier (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
      * @param[in]      act_info         (Optional) Activation layer information in case of a fused activation.
      * @param[in]      dilation         (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
+     *
+     * @note: For in-place support, please check @ref CLDepthwiseConvolutionLayerNativeKernel
      */
-    void configure(ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info, unsigned int depth_multiplier = 1,
-                   ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U));
+    void configure(const CLCompileContext &compile_context, ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
+                   unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U));
+
     /** Initialize the function's source, destination, weights and convolution information.
      *
-     * @param[in]      compile_context  The compile context to be used.
-     * @param[in, out] input            Source tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/FP16/FP32. Data layout supported: NHWC, NCHW
-     * @param[in]      weights          Weights tensor. These are 3D tensors with shape [kernel_x, kernel_y, IFM].
-     *                                  Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
-     * @param[in]      biases           Biases tensor. A 1D tensor with shape [IFM]. Must be nullptr if not needed.
-     *                                  Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out]     output           Destination tensor. Data type supported: same as @p input.
-     * @param[in]      conv_info        Padding and stride information to use for the convolution.
-     * @param[in]      depth_multiplier (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]      act_info         (Optional) Activation layer information in case of a fused activation.
-     * @param[in]      dilation         (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
+     * Similar to @ref CLDepthwiseConvolutionLayer::configure()
      */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
+    void configure(ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
                    unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U));
 
     /** Static function to check if given info will lead to a valid configuration of @ref CLDepthwiseConvolutionLayer
      *
-     * @param[in] input            Source tensor info. Data type supported: QASYMM8/QASYMM8_SIGNED/FP16/FP32. Data layout supported: NHWC, NCHW
-     * @param[in] weights          Weights tensor info. These are 3D tensors with shape [kernel_x, kernel_y, IFM].
-     *                             Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
-     * @param[in] biases           Biases tensor info. A 1D tensor with shape [IFM]. Must be nullptr if not needed.
-     *                             Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in] output           Destination tensor. Data type supported: same as @p input.
-     * @param[in] conv_info        Padding and stride information to use for the convolution.
-     * @param[in] depth_multiplier (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in] act_info         (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU for 3x3 QASYMM8 supported.
-     * @param[in] dilation         (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
+     * Similar to @ref CLDepthwiseConvolutionLayer::configure()
      *
      * @return a status
      */