Installing Debian 11.11 on and Nvidia DGX V100 Station

Ask for help with issues regarding the Installations of the Debian O/S.
Post Reply
Message
Author
Scott_Best
Posts: 1
Joined: 2024-10-16 10:11
LinkedIN: https://www.linkedin

Installing Debian 11.11 on and Nvidia DGX V100 Station

#1 Post by Scott_Best »

I have an Nvidia DGX V100 Station that contains a Xeon processor with four V100 GPU cards, which one V100 also working as the video interface. Nvidia uses Ubuntu as the foundation for the Nvidia DGX OS, and the system runs without any problems using Ubuntu 24.01 and 24.1. My client owns the DGX Station, and we are working on a large MATLAB project with Software Defined Radios (SDRs). MATLAB 2024a has numerous bugs, which renders it unusable in Ubuntu, which MathWorks Tech Support confirmed. They are aware of the bugs, but do not have a timeframe for fixing them. MathWorks uses Debian as their Linux system for this reason, so they asked that Debian 11.11 be installed on the DGX Station so the project can move forward using MATLAB 2024a.

I downloaded "debian-11.11.0-amd64-DVD-1.iso" and moved the installation files to a Flash Drive using Rufus 4.5. The installation proceeded as follows:

1. Debian 11.11 Installation Steps
2. Advanced Options Selected
3. Expert Install Selected
4. ... intermediate steps
5. generic: include all available drivers
6. Network Mirror - YES
7. Non-Free Software - YES
7.a. /etc/apt/sources.list - YES, but I did not find this list on the Flash Drive, so do I need to edit this file?
8. Debian Desktop Environment + GNOME + Mate (Mate is used by MathWork personnel for running MATLAB.) + Standard System Utilities
9. Default Display Manager - gdm3
10. Grub Boot Loader

The installation is completed without any warnings or errors, and the DGX Station is rebooted to start Debian.

Debian reboots, but defaults to a black screen with the cursor shown in the top left corner. Many other people using Debian have experienced this problem, but their solutions to fix this problem are not for installing Debian on a system. Therefore, can anyone provide some guidance or step-by-step instructions for me to follow for fixing this problem on the existing installation, or with a new installation of Debian? I have been working on this for two days without any success, so your assistance with this matter would be greatly appreciated.

Thanks in advance for your assistance with this installation problem with Debian 11.11,

Scott

CwF
Global Moderator
Global Moderator
Posts: 3077
Joined: 2018-06-20 15:16
Location: Colorado
Has thanked: 63 times
Been thanked: 254 times

Re: Installing Debian 11.11 on and Nvidia DGX V100 Station

#2 Post by CwF »

I would step back and separate out the issues. The blank screen is one issue that is maybe not related to the function of the target softwares. Meaning, it would be a waste to solve the primary display issue to then find other issues. If possible I would temporarily treat the box as headless and try to verify all the extra requirements are working. If you can ssh in, forward X, and run things that way and it all checks out, then the primary display issue may be solved with an alternative gpu for the purpose, if necessary and possible.

Since there are many details to consider I'll stop there.
Mottainai

Post Reply