Discovery Cluster
Research Computing provides access to Discovery Cluster GPUs via HPC.
The Discovery Cluster refers to a group of high-powered Linux-based computers, totalling 128 nodes, 6712 CPU cores, 54.7TB of memory, and more than 2.8 PB of disk space. This world-class concentration of computing resources allows researchers to meet a wide range of needs. Research Computing support is available to support researchers in installing software; programming, debugging, and optimizing code; supporting infrastructure (hardware/networking) needs; and providing access to job scheduling tools using Slurm.
The “Free tier” allows access to high-performance, large-memory computing systems on a shared basis. The “Paid tier” provides access to GPU shares (currently V100s and A100s) among a shared community of users.
There are several helpful KB articles available on services.dartmouth.edu: