Seneca Cluster Documentation
The Seneca cluster is a High-Performance Computing (HPC) system jointly managed by Cornell’s Center for Advanced Computing (CAC), Weill Cornell Medicine Qatar Advanced Computing team (AdvC), and the Weill Cornell Medicine Scientific Computing Unit (SCU). It is available for research users with non-sensitive data.
The cluster features modern CPUs, NVIDIA H100 and A100 GPUs, a BeeGFS parallel file system, and a 400 Gb NDR InfiniBand interconnect. It runs on Rocky Linux 9 and uses Slurm 25.05.3 for job scheduling.
Hardware and Networking
Servers:
GPU Specifications
To access the GPUs, submit a Slurm job to the cac_gpu partition using the --gres option.
Note: Compared to the H100 PCIe, the H100 SXM offers slightly more compute power, significantly faster memory, and high-speed NVLink connections between GPUs on the same node.
Storage
Note: Storage systems on the Seneca cluster are not backed up.
- User Home Directory: Each user has a home directory with a 100 GB default quota, accessed via
~. - BeeGFS Parallel File System: Each project has a directory at
/mnt/beegfs/<institution>/<project name>readable/writable by project members. Institutions includeithaca,qatar, andweill. - Local Scratch: Mounted on
/tmpon all nodes.
In addition to scp/sftp, storage is accessible via the Seneca Cluster Globus collection.
Request Access
For Weill Cornell Medicine Principal Investigators (PIs) requesting access to the Seneca cluster through the Scientific Computing Unit (SCU), please note that provisioning requires an active fund number and the completion of a data compliance review prior to resource allocation.