So, i have a Tesla M1060 installed on a copmputer and i installed all the drivers and libraries needed, i compiled the GPU Computing SDK sample code and it runs… BUT when i try to run it as regular used (not root) it hangs… any idea?
My machine is a Xeon double core and Debian Stable.
So, i have a Tesla M1060 installed on a copmputer and i installed all the drivers and libraries needed, i compiled the GPU Computing SDK sample code and it runs… BUT when i try to run it as regular used (not root) it hangs… any idea?
My machine is a Xeon double core and Debian Stable.
What are the ownership and permission settings for the /dev/nvidia* files? Usually when people have a problem with the non-root user, it is because their regular user does not have read/write access to the /dev/nvida* files. Normally this results in an error message saying that no devices were found, so I’m a little surprised that your code just hangs.
What are the ownership and permission settings for the /dev/nvidia* files? Usually when people have a problem with the non-root user, it is because their regular user does not have read/write access to the /dev/nvida* files. Normally this results in an error message saying that no devices were found, so I’m a little surprised that your code just hangs.
No, those permissions should allow non-root users to use the devices. I’ve never seen an actual hang before. Can you say what Linux distribution, NVIDIA driver version, and CUDA toolkit version you are using?
At least under OpenSuSE it usually suffices, to add the users to group video. rw-rw-rw- when set by hand usually gets reset upon the next reboot unless explicitely changed in some obscure permissions files in /etc.