Currently, Graphics Processing Units (GPUs) are taking the spotlight in high performance computing. Researchers in many scientific domains have found that GPUs can accelerate their research codes and reduce the "time to science". This acceleration has also permitted researchers to tackle larger and more complex problems. GPU hackathons provide a one week hands-on experience where vendors and experts guide researchers towards accelerating their codes with GPUs.
Learn more at https://gpuhackathon.com
Fluid Numerics is helping to organize and moderate the NCSA and Brookhaven GPU Hackathons. In support of these hackathons, we have published the first version of attendee and mentor guidelines. Additionally, Fluid Numerics has set up access to a virtual GPU cluster on Google Cloud Platform for the Brookhaven event. This cluster consists of 16 compute nodes, each with a Tesla V100 GPU, and a standard set of HPC packages with GCC and PGI compilers.
Thanks for attending the 2019 Santa Fe GPU Hackathon!
We had a great time in Santa Fe, NM with teams from all over the country. Our heterogeneous high performance cloud cluster realized some great test runs and was able to grow to assist even more educators, researchers and scientists. If you missed the event and are interested in gaining an understanding about how our cluster on Google Cloud Platform(GCP), a multi-node "bare-metal" accelerated computing cluster, could help your team just ask. Keep in touch by subscribing to the Fluid Numerics Journal for exciting updates about hack-a-thons and our advancements within scientific computing on the cloud!