Many scientific and analytic applications can benefit from the acceleration available by running on GPUs in addition to, or instead of, just the CPUs available in compute nodes. YCRC’s clusters have a variety of GPUs available for use via the gpu and gpu_devel partitions on Grace and McCleary, and the gpu partition on Milgram. More GPUs, and additional types, are available via the scavenge_gpu partition on Grace and McCleary and the scavenge partition on Milgram. Please be sure to request appropriate GPUs be allocated to your jobs. More information can be found in the YCRC’s HPC Clusters User Guide.
Please contact hpc@yale.edu with any questions.