C2050 on a network Access problems

Accelerated Computing CUDA CUDA Programming and Performance
1

Dear all,

We have two C2050 GPUs on one of the machines on our group network. Recently, no users except me can detect the GPUs when they login to the machine.

Since we are presently between sys-admins, I was hoping someone in this group can point me in the right direction. Which config file/s should I be looking at to rectify this?

Thanks in advance for any suggestions,

Prashanth

2

Dear all,

We have two C2050 GPUs on one of the machines on our group network. Recently, no users except me can detect the GPUs when they login to the machine.

Since we are presently between sys-admins, I was hoping someone in this group can point me in the right direction. Which config file/s should I be looking at to rectify this?

Thanks in advance for any suggestions,

Prashanth

3

Have a look at whether the device files /dev/nvidia0 and /dev/nvidia1 are present and what their permissions are. If they don’t exist then they need to be created. There is a discussion in the release notes for the cuda Linux toolkit which describe how to do this. Otherwise change the permissions so that users have both read and wrote permission to the devices.

4

Have a look at whether the device files /dev/nvidia0 and /dev/nvidia1 are present and what their permissions are. If they don’t exist then they need to be created. There is a discussion in the release notes for the cuda Linux toolkit which describe how to do this. Otherwise change the permissions so that users have both read and wrote permission to the devices.

5

Thanks very much – that worked.

-Prashanth

6

Thanks very much – that worked.

-Prashanth