tkonolige commented on pull request #8119: URL: https://github.com/apache/tvm/pull/8119#issuecomment-847182571
It seems like the host version of CUDA has to match what is running in the containers (supposedly this is not the case, but making the versions match fixed all the issues we were seeing). If we want to test against CUDA 10 and 11, were going to have to partition the GPU nodes. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
