Shantenu, Matteo: can you advise on point 1? I do not know at what scale RCT has been used on Summit in context of Covid (or in other contexts). Also on point 3: what are the task sizes for DeepdriveMD? I assume the smallest is 1 task per GPU - is that correct?
Thanks, Andre. ---------- Forwarded message --------- From: Pankajakshan, Ramesh <pankajaksh...@llnl.gov> Date: Mon, Apr 20, 2020 at 7:16 PM Subject: Running at scale To: mikhail.a.ti...@gmail.com <mikhail.a.ti...@gmail.com>, an...@merzky.net <an...@merzky.net> Andre, Mikhail; A couple of questions about your framework: What is the largest number of nodes that you have used for a job on Summit ? Do you use python from a virtual env ? Typically how many python instances do you need per CORAL node ? Thanks Ramesh _______________________________________________ mtt-users mailing list mtt-users@lists.open-mpi.org https://lists.open-mpi.org/mailman/listinfo/mtt-users