Hi,
Since we don’t own OpenCV implementation, you can check this issue with OpenCV developer to get more information.
Based on their code, it seems that they implement convolution through cufft rather than cudnn .
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
show original
For slow cuDNN issue, this is a known regression from cuDNN v8.https://forums.developer.nvidia.com/t/darknet-slower-using-jetpack-4-4-cudnn-8-0-0-cuda-10-2-than-jetpack-4-3-cudnn-7-6-3-cuda-10-0/
Thanks.
1 Like