Are you new to ARCC and HPC?
If you are just starting out, our team can assist you in getting up to speed using our HPC resources.
We recommend prospective users follow the directions here to get access to ARCC Cluster Computing to get an account and start using our cluster resources.
Request a Project
New Hardware Updates on Medicinebow
Storage
Medicinebow uses the VAST data filesystem configured with over 3 petabytes of NVMe storage, including advanced block-based data deduplication expected to provide over 6PB of overall research data storage capcity.
About VastCompute Nodes
ARCC adds 25 new 2x 48-core/96-thread 4th Gen AMD EPYC 9454 nodes to the existing 300 heterogeneous nodes from our Beartooth cluster.
GPU Nodes
Medicinebow includes the integration of several specialty GPU nodes including 8 x H100, 8 x L40S, and 8 x A30 GPUs available for researchers to leverage powerful GPU hardware towards deep learning, complex modeling, or otherwise intensive miscellaneous tasks. More detailed information about Medicinebow nodes is available on our Medicinebow Hardware Summary.
Improvements to Infrastructure
The Medicinebow cluster features several newer GPUs running intense computations. ARCC has therefore installed more powerful infrastructure within the UW IT Data Center to maintain these nodes including liquid cooled racks and 1TBPS RoCEv2 networking.