I am running Windows 8.1 (64bit) with Matlab 2016b and Cuda 8.0. Up to this point I had been using a GeForce GTX TITAN X and everything was working correctly. I just added a second GPU in my system (TITAN X (Pascal)) and I get the "Attempt to execute SCRIPT vl_nnconv as a function" error when trying to use both with cnn_train.
If I only use each card individually the code is executing as expected without any errors (although the Pascal card takes longer to start, which is a known issue accordin to Mathworks). The problem only appears if I attempt to use both GPUs at the same time.
Do you have any suggestions? I tried searching for similar issues online but I could not find anything relevant.
I am running Windows 8.1 (64bit) with Matlab 2016b and Cuda 8.0. Up to this point I had been using a GeForce GTX TITAN X and everything was working correctly. I just added a second GPU in my system (TITAN X (Pascal)) and I get the "Attempt to execute SCRIPT vl_nnconv as a function" error when trying to use both with cnn_train.
If I only use each card individually the code is executing as expected without any errors (although the Pascal card takes longer to start, which is a known issue accordin to Mathworks). The problem only appears if I attempt to use both GPUs at the same time.
Do you have any suggestions? I tried searching for similar issues online but I could not find anything relevant.
Regards, Anestis