Kind of hoping we get to use these huge gpu clusters to run massive molecular dynamics simulations.

Tim Duignan What would be a kind of PoC experiment that you think would be really cool, and how many GPU-hours? If interested, we or some group could consider applying to one of:
new.nsf.gov/focus-areas/ar…
doeleadershipcomputing.org/about/
access-ci.org
Non-US collabs are generally allowed.

Ken Chiu Yeah great question and idea thanks I will think some more about it

Tim Duignan NSF ACCESS is the easiest to get. Below is a chart of the different allocations. Exchange rate is about 50 credits per GPU-hour. There is no charge, but they like to see that it is for funded research, but pretty much any grant will do. 1/3

Tim Duignan If other non-US researchers are interested, I'm happy to act as the US collaborator. Or, if there are US researchers that are not very familiar with running on large Linux clusters, I can also help out there. 2/3

Tim Duignan In 2023, there were two materials/molecules related submissions to the GB prize: sc23.supercomputing.org/2023/08/a-look…. 3/3