I would like to have CUDA 8 and 9 installed in parallel and both be functional, but I am not sure if this is possible because of the Nvidia driver they both need seems to be different. Is this at all possible?
Yes, it’s possible. the CUDA 9 driver will work with CUDA 8.
Hi, are you positive I tried to use NVIDIA-Linux-x86_64-390.42 (which CUDA 9 requires), but it doesn’t work with CUDA 8.
It works with CUDA 8.
I’m comming across the following error whenever I’m trying to use cuda on Ubuntu 16.04, CUDA 9, CUDNN 5.0.4
libcublas.so.9.0: cannot open shared object file: No such file or directory
I did follow every step specified for installing tensorflow-gpu, I’m getting this error which is probably I’m encountering due to some issues in cuda installation