Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.

NCSA's Innovative Systems Lab hosts the ISL-cluster as a testbed for new HPC servers and services, combining best practices in cluster management with testing on new methods of cluster access, management, services and deployment. The cluster is active in testing and new operational services, so uptime and availability are not guaranteed. Please ensure that you keep a copy of all important data elsewhere.

In publications and presentations that use results obtained on this system, please include the following acknowledgement: “This work utilizes resources supported by University of Illinois at Urbana-Champaign”.

Innovative Systems Lab (ISL) cluster

Host name:


  • 4x GPU compute nodes
    • Dell PowerEdge R730
    • 2x Intel CPUs
    • 2x NVIDIA V100 GPUs
  • Dell testbed nodes
    • AMD Genoa CPU node (high core/high ram)
    • AMD Genoa GPU node w/ MI2xx (AMD GPU testbed)
    • Intel Sapphire Rapids CPU node (high core count)
    • Intel Sapphire Rapids HBM CPU node (high core count)



To request access: contact Dan Lapine at

To report problems: email us with ISL-cluster in the subject

Real-time Dashboards: (work in progress)

Quick start guide:

To connect to the cluster:

Code Block
ssh <username> 

To submit interactive job:

Code Block
srun --pty -p cpu -- bash

To submit interactive job to the himem node:

Code Block
srun --pty -p cpu -w isl-cpu-genoa01 -- bash

To submit a batch job:

Code Block

Job Queue time limits:

  • "gpu" and "cpu" queues:  24 hours


  • shared /home
    • nfs served from isl-login01
  • shared /project space
    • lustre on flash storage served via Infiniband
  • Taiga access on isl-cluster01
    • nfs served via 10Gbe link

Resource limits:

  • 5GB quota for /home
  • 1TB quota for /projects

To see CLI scheduler status:

Code Block

Main -> Systems -> Isl cluster

Contact us

Request access to this system:

Contact ISL staff: Email Address

Visit: NCSA, room 3006H