Ask Your Question
1

Building OpenCV 4 with CUDA

asked 2019-10-21 13:36:10 -0500

Erik Langskjegg gravatar image

updated 2019-10-21 13:37:42 -0500

I am so exited for the merging of the GSoC project enabling CUDA support in OpenCV!

However I am uncertain about the build procedure shown in this gist. I think I understand setting cmake build options like this:

cmake -D CMAKE_BUILD_TYPE=RELEASE \
  -D CMAKE_INSTALL_PREFIX=/usr/local \
  -D WITH_CUDA=ON \
  -D WITH_CUDNN=ON \

Where do I set the "following option"

OPENCV_DNN_CUDA

then?

edit retag flag offensive close merge delete

Comments

1

FYI: OpenCV has had CUDA support for a few years. The linked PR only adds CUDA support to the OpenCV DNN module.

Yashas gravatar imageYashas ( 2019-10-27 13:06:34 -0500 )edit

I understand, sorry for my lack of precision. It looks like your CUDA implementation will make a fine impact on the DNN module though?

Erik Langskjegg gravatar imageErik Langskjegg ( 2019-10-27 13:15:10 -0500 )edit
1

Yes, it adds a new backend to the DNN module which accelerates DNN inference on CUDA capable cards dramatically.

Yashas gravatar imageYashas ( 2019-10-27 23:39:59 -0500 )edit

0 answers

Sort by ยป oldest newest most voted
4

answered 2019-10-27 12:52:07 -0500

Yashas gravatar image

updated 2019-10-27 12:53:27 -0500

OPENCV_DNN_CUDA is also a CMake option. You can set it using -DOPENCV_DNN_CUDA=ON.

edit flag offensive delete link more

Comments

thanks so much, I can't wait to try out your Google project!

Erik Langskjegg gravatar imageErik Langskjegg ( 2019-10-27 12:56:34 -0500 )edit
Login/Signup to Answer

Question Tools

1 follower

Stats

Asked: 2019-10-21 13:36:10 -0500

Seen: 121 times

Last updated: Oct 27