Case Study: SUNY Downstate Medical Center cuts compute time from days to hours with Google for Education

SUNY Downstate Medical Center deploys Slurm integration with GCP, cutting compute time from five days to one hour

SUNY Downstate Medical Center, part of the State University of New York system, needed more computing power to support research across its five colleges and hospital. Facing outdated on-premises equipment and growing demand from faculty and researchers at the Neurosim Lab, the center looked to Google for Education and Google Cloud Platform to enable scalable, high-performance computing for complex brain-circuit simulations.

Google for Education helped SUNY Downstate deploy Google Cloud Platform Compute Engine with Preemptible VMs and Slurm integration to auto-scale its HPC workloads. The result was a dramatic speedup: simulations that previously took three to four days now run in three to four hours, and one workload was reduced from five days to about one hour. The team can also spin up as many as 2,300 nodes in minutes, cut costs by about 4x with Preemptible VM instances, and run far more hypotheses faster.

SUNY Downstate Medical Center

Roy Sookhoo

Chief Information Officer