On Thu, Jun 26, 2014 at 7:36 AM, Prentice Bisbal <prentice.bis...@rutgers.edu> wrote: > I disagree with needed separate images in a VM/Docker/etc. environment, > maybe if your doing system-level research, but most HPC users work > exclusively in the user-space,and just want to run their MATLAB/NAMD/LAMMPS, > whatever job. In this case, just installing user-space applications and > libraries in a different path from the distro-supplied versions is adequate. > Modifying PATH and a few other environment variables that most users can > handle is all you need modules, lmod, softenv and other utilities make that > even easier for users.
+1 on this. If users start requesting kernel modifications or low-level system tuning, it's likely the sign that a shared system is not a good fit for their needs, and that they would probably benefit from using and managing their own system. To me, the main interest of a shared system is that users can just log in, put together a submission script using the already installed and configured software they need, and go off doing science. All the hardware, low-level tuning and system administration is taken care of for them so they don't need to waste time managing these things. And they can make a better use of their time, focused on their simulations. > If you're doing lower-level research involving kernel modules or kernel > tuning, yes, you will need VMs or something. But this is usually 'research', > not 'production' HPC. Yup. > This is so true. If you are a roadblock to your users, they will find a way > around you. And make your life hell. They have plenty of ways. :) Cheers, -- Kilian _______________________________________________ Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
