Thanks for the suggestions – some interesting links to read. We already have an option to run graphical sessions from the head node using Apache Guacamole which works well, but that still requires X11 to talk to the compute nodes. We don’t have a full desktop stack on the compute nodes and just want to run individual applications.
The issue isn’t network bandwidth – I can launch two graphical instances, one via ssh and the other via srun at the same time on the same compute node and ssh is great but srun is terrible. We do route through the head node (the compute nodes aren’t directly addressable) but the overall traffic on the head node is pretty modest.
I’m not really sure how the X11 plugin works (assuming it’s not just doing ssh X tunnelling) to try to think what else could be limiting here.
Simon.
From: Jason Simms <jsimms1@swarthmore.edu>
Sent: 06 June 2025 13:52
To: Hadrian Djohari <hxd58@case.edu>
Cc: John Hearns <hearnsj@gmail.com>; Simon Andrews <simon.andrews@babraham.ac.uk>; slurm-users@lists.schedmd.com
Subject: Re: [slurm-users] Re: X11 performance terrible using plugin
CAUTION: This email originated outside of the Organisation. Please help to keep us safe and do not click links or open attachments unless you recognise the sender and know the content is safe.
It may or may not be an appropriate solution for your use cases, but I second using Open OnDemand and its virtual desktop. It is FAR more performant than X11 through Slurm/SSH.
Jason L. Simms, Ph.D., M.P.H.
Research Computing Manager
Swarthmore College
Information Technology Services(610) 328-8102
On Fri, Jun 6, 2025 at 8:40 AM Hadrian Djohari via slurm-users <slurm-users@lists.schedmd.com> wrote:
Or use Open OnDemand platform for the interactive Desktop.
On Fri, Jun 6, 2025 at 8:37 AM John Hearns via slurm-users <slurm-users@lists.schedmd.com> wrote:
Simon, I have had success in the past by using NICE DCV (now owned by AWS but you can get licenses for on prem)
https://www.ni-sp.com/products/nice-dcv
An alternative would be VirtualGL
Altair Access (though more likely to work with PBS!) https://altair.com/access\
On Fri, 6 Jun 2025 at 10:42, Simon Andrews via slurm-users <slurm-users@lists.schedmd.com> wrote:
On our cluster we’ve noticed that if we use the native x11 slurm plugin (PrologFlags=x11) then X applications work, but are really slow and unresponsive. Even opening menus on graphical application is painfully slow.
On the same system if I do a direct ssh connection with ssh -YC from the head node the same applications are quick and responsive.
Any suggestions for what might be causing this, and how I can get the native x11 to have the same responsiveness as a direct ssh connection?
Many thanks
Simon.
--
slurm-users mailing list -- slurm-users@lists.schedmd.com
To unsubscribe send an email to slurm-users-leave@lists.schedmd.com
--
slurm-users mailing list -- slurm-users@lists.schedmd.com
To unsubscribe send an email to slurm-users-leave@lists.schedmd.com
--Hadrian Djohari
Director of Advanced Research Computing, [U]Tech
Case Western Reserve University
(W): 216-368-0395
(M): 216-798-7490
--
slurm-users mailing list -- slurm-users@lists.schedmd.com
To unsubscribe send an email to slurm-users-leave@lists.schedmd.com
This email has been scanned for spam & viruses. If you believe this email should have been stopped by our filters, click here to report it.
--
slurm-users mailing list -- slurm-users@lists.schedmd.com
To unsubscribe send an email to slurm-users-leave@lists.schedmd.com