In late September I attended a conference of the National Science Foundation’s Campus Cyberinfrastructure program, on behalf of the University of Colorado Boulder and our research computing team. This year’s event took place in Columbus, Ohio, in a nice convention center.
Visiting Columbus constituted a bit of a return to my past work and home at Earlham College just a few hours from there. I didn’t have time for a visit, but it was a bit nostalgic just the same.
It was also cloudy and cool on arrival, a reprieve from high temperatures we were still experiencing in northern Colorado at the time.
I spoke for about ten minutes on an NSF award we received to expand the Alpine cluster to support more users from the Rocky Mountain Advanced Computing Consortium (RMACC). This is a project that got picked up by HPCWire using text I drafted alongside RMACC’s executive director. That’s fun!
Some key takeaways:
- Computing needs in higher education to continue to grow. This is certainly true of research but it also applies to courses, as the large language model (LLM, colloquially “AI”) surge brings high compute and data storage requirements.
- This is perhaps obvious but I think it’s worth saying explicitly: needs and approaches vary widely. Some institutions, like CU, run centralized resources that everyone on campus (as well as external contributors) can use. Others build out dedicated HPC clusters customized for a particular domain science. Still others exclusively run a condo (researcher buy-in) model.
- The campus cyberinfrastructure community is very open and friendly. This was my first conference in this program and I never felt out of place.
- Researcher trust is central to success in building out campus cyberinfrastructure.
- Typing “cyberinfrastructure” over and over again is a pain, which is why they abbreviate it to CC*. 🙂
Ultimately I had a good time. I’m looking forward to implementing this project.