MATLAB & Cuda 10

5 views (last 30 days)
Pol Henarejos
Pol Henarejos on 26 Nov 2018
Answered: Joss Knight on 30 Nov 2018
Dear all,
I have a NVIDIA Turing GPU card and it requires Cuda 10. I see that Matlab 2018b uses Cuda toolkit 9.1. How could I use Matlab 2018b with my Turing card? Is there a way to tell Matlab to point to another Cuda version?
Thanks.

Answers (2)

Andrew Borghesani
Andrew Borghesani on 29 Nov 2018
Have you tried to actually use it? CUDA is backwards compatible and should work, although the libraries may need to be recompiled which may take some time. Please see here for more detail regarding specific CUDA support by GPU archtecture: https://www.mathworks.com/help/distcomp/gpu-support-by-release.html

Joss Knight
Joss Knight on 30 Nov 2018
It works, although it will erroneously warn that JIT compilation is required (it isn't). Also, there is a known bug with Deep Learning which can be worked-around by running a simple function and ignoring the error
try
nnet.internal.cnngpu.reluForward(1);
catch ME
end

Tags

Products


Release

R2018b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!