I’m on a Centos 7 machine trying to get Cuda 9.2 to run. When I run the deviceQuery, I get “CUDA driver version is insufficient for CUDA runtime version”, and when I run a simple program I get a segmentation fault.
After trying various things all morning, I uninstalled the previous version (installed via RPM), reinstalled an NVidia driver, then installed Cuda 9.2 using the run file. My Xserver wouldn’t start, so I downloaded and installed the latest NVidia driver (390.59), and the Xserver starts, but Cuda won’t work. I saw a post that said "use the driver that came with Cuda 9.2, but I have two problems with that.
First, HOW do I use that driver?
Second, will it work, because when I installed Cuda it broke my Xserver (or something–all I know is that the actual machine came up with either a blank screen or an error screen), and I assume it was using that driver. And is there a difference between using that driver for Cuda programs and using that driver for the Xserver?
Please help. I’m lost.