It’s only 25 miles from NCAR to NREL, so I have time to gas up, get a sandwich, and even catch a car wash before my tour stop at NREL. Feeling refreshed, I pull up to their Golden, CO, campus and start the interview with Ray Grout, the director of NREL’s Computational Science Center.
NREL is all about energy efficiency and renewable energy, full stop. To support their mission, they primarily rely on Eagle, a 8 petaflops supercomputer, plus some assorted smaller clusters. They’re eagerly anticipating the arrival of Kestrel, which will weigh in at around 44 petaflops. This will be more heavily accelerated than Eagle, sporting large numbers of GPUs. Which GPUs and how many will probably be released during SC22.
In the interview, we discuss how the lab is embracing AI and more specifically, machine learning. In keeping with the lab’s mission, they’re also working to implement green AI, with the goal of reducing the energy used by AI/machine learning computation.
We had a nice segue into the trade-offs between specialized vs. general purpose computing hardware. We also touched on a very recent meme I’ve been hearing recently – where scientists and researchers are expressing interest in how much energy their specific runs will consume. As I’ve mentioned before, this point has come up repeatedly in my recent travels.
We had a fun and thoughtful interview and I really appreciated them taking the time to arrange this tour stop. We also appreciate our sponsors who have made this possible – HPE, Cornelis Networks, GigaIO, and Dell Technologies. Plus our pals at HPCwire, of course.
Next stop: Los Alamos National Laboratory in lovely Los Alamos, NM, which is about 350 miles down the road.
Stay tuned for more updates!