High Performance Computing (HPC)¶
Welcome to HPC at New Jersey Institute of Technology (NJIT).
NJIT provides High Performance Computing resources to support scientific computing for faculty and students. These resources include CPU nodes, GPU nodes, parallel storage, high speed, low latency Infiniband networking and a fully optimized scientific software stack.
-
Virtual Tour of NJIT Data Center¶
Wulver is built through a partnership with DataBank, which is live in DataBank’s Piscataway, N.J. data center (EWR2) and will support NJIT’s research efforts. This infrastructure will bolster NJIT’s research initiatives. You can access the 3D virtual tour of HPC data center below:
-
HPC Highlights¶
Contact Us
To create a ticket or request for software installation, visit Contact Us.
Clusters
See the list of availble clusters, their specifications, and how to access them in Clusters.
Software Modules
See Software Modules for list of software packages installed on our cluster.
Lochness to Wulver Migration
For details in migrating the Lochness users to Wulver, please visit Migration.
Cluster Maintenance Updates
To see the latest maintenance updates on NJIT cluster, please visit Cluster Maintenance.
HPC Events
Check the upcoming events hosted by HPC team and register from HPC Events.
HPC Trainings and Webinars
Make sure to stay up-to-date with the latest webinar sessions on HPC by visiting the HPC Training.
-
HPC latest News!¶
HPC Fall Events
ARCS HPC is hosting the following events this fall. Please register for the events you plan to attend.
- SLURM Batch System Basics
- Introduction to Containers on Wulver
- Job Arrays and Advanced Submission Techniques for HPC
For dates and registration, visit Fall Events.
Wulver Scheduled Maintenance
Wulver will be out of service for maintenance once a month for updates, repairs, and upgrades. The schedule is 9 a.m. to 9 p.m. the second Tuesday of every month. During the maintenance period, the logins will be disabled and the jobs that do not end before the maintenance window begins will be held until the maintenance is complete and the cluster is returned to production. For example, if you submit a job the day before maintenance, your job will enter a pending state (you will see job status
PD
when usingsqueue -u $LOGNAME
). You can either adjust the walltime or wait until maintenance ends. Please stay informed about maintenance updates at Cluster Maintenance Updates and News.