From: Matthew Kenney (
Date: Tue Jul 10 2018 - 16:47:00 CDT

Hi Everyone,

This message continues the discussion listed under the previous subject:
"Issue with NVIDIA GPU Cloud VMD Container."

Since my previous post, I have been able to use a multi-stage docker build
to extend the NVIDIA vmd image to include glx support. Further, I have been
able to extend the image to enable the container to access the host's
display. This image, and its corresponding README file is hosted on github:

Within the container built from my customized image, I was able to both
successfully run glxgears, and use text-based vmd without issue. This
verifies that the container both:

1) Works with and properly displays glx applications
2) Continues to run vmd properly from the text-based vmd version contained
in the image

The image I currently have, surprisingly, is just about 0.1 GB larger than
the original vmd image, indicating that adding glx functionality did not
bloat the image size too excessively.

The last step to getting vmd to work visually from within the container is
to download and compile a visual version of vmd from within the container.
This has proved difficult for me to do, however, because the original vmd
image from the NVIDIA GPU cloud does not have a publicly available
dockerfile, and thus I cannot see how the text-based vmd was downloaded in
the first place.

If anyone has a way around this issue, please let me know. Alternatively,
since you worked on this project, john, if you could direct me on what
steps I might take from here, or check out my github page and see if you
would like to collaborate, that would be very much appreciated. Thank you
for your time!


This message contains information which may be confidential and/or 
protected by attorney-client privilege.  Unless you are the addressee, you 
may not use, copy or disclose to anyone this message or any information 
contained in this message.  If you have received this message in error, 
please send me an email and delete this message.   Thank you.