Tesla P4 remote rendering, VNC and Centos 7

Hi all,

I am using a Tesla P4 card in a Centos 7 operating system. I have installed “nvidia-diag-driver-local-repo-rhel7-390.46-1.0-1.x86_64.rpm” following installation instructions in the download page. I had blacklisted nouveau drivers and verified them by checking lsmod| grep nouveau. The system is a Dell Poweredge 730 and has a “VGA compatible controller: Matrox Electronics Systems Ltd. G200eR2”.

The issue is that when I VNC into the machine and launch nvidia X server settings a popup appears saying “You do not appear to be using the nvidia X server” and then the screen goes blank.

Could someone please help me with this issue?

I am attaching the Xorg.0.log and nvidia bug report.


Xorg.0.log (24 KB)
nvidia-bug-report.log.gz (141 KB)

Logs look ok. Are you sure you’re actually connecting to the nvidia xserver e.g. through x11vnc or are you using xvncserver? That’s a plain software xserver.

Thanks generix. I understood the difference only after I posted this. Yes, I am able to connect via x0vncserver to the display 0.
Is there a way if I wanted hardware acceleration on other vncservers also, without using virtualGL.


You would either have to start a separate xserver on the nvidia gpu for each vnc client you want to use or set up virtualgl.

I use tigervnc server, I am not sure how to start a separate x server for each client. Could you please point me to some reference or help me with this.