COMPMID-1413 - Improve the performance of GEMMLowp with 8 bit dot product on OpenCL
COMPMID-1424 - Add dot product support for CLDepthwise QASYMM8 3x3 NHWC non-unit stride
With this patch we are able to improve the performance of MobileNet v1-qasymm8 by 37 %
Tried to use the dot product instruction in CLDepthwise QASYMM8 3x3 NHWC non-unit stride
but I have not seen any benefit (maybe because we have few arithemtic operation and we
do not have more load instructions). However Depthwise convolution has been improved by
30%
Change-Id: Id768a99c2e53a04276707e427af5d0ec93419ada
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/155082
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h
index 1206206..72373b5 100644
--- a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h
+++ b/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.h
@@ -67,25 +67,22 @@
* @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8
* @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
* Along with @p min, this value can be used to implement "rectified linear unit" activation functions
- * @param[in] output_3d_depth (Optional) Depth of output in 3D (Defaults to 1)
*/
void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, int result_fixedpoint_multiplier, int result_shift, int result_offset_after_shift,
- int min = 0, int max = 0, unsigned int output_3d_depth = 1);
+ int min = 0, int max = 0);
/** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel
*
- * @param[in] input Input tensor. Data type supported: S32
- * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
- * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
- * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8
- * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8
- * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
+ * @param[in] input Input tensor. Data type supported: S32
+ * @param[in] bias Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
+ * Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
+ * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8
+ * @param[in] min (Optional) Min value used to saturate down the output result before converting back to QASYMM8
+ * @param[in] max (Optional) Max value used to saturate up the output result before converting back to QASYMM8,
* Along with @p min, this value can be used to implement "rectified linear unit" activation functions
- * @param[in] output_3d_depth (Optional) Depth of output in 3D (Defaults to 1)
*
* @return a status
*/
- static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output,
- int min = 0, int max = 0, unsigned int output_3d_depth = 1);
+ static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, int min = 0, int max = 0);
// Inherited methods overridden:
void run(const Window &window, cl::CommandQueue &queue) override;
@@ -94,7 +91,6 @@
const ICLTensor *_input;
const ICLTensor *_bias;
ICLTensor *_output;
- bool _reinterpret_as_3d;
};
} // namespace arm_compute
#endif /* __ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32TOUINT8SCALEBYFIXEDPOINTKERNEL_H__ */