Setting up the account

Access details will be provided to the team leader.

Teams	Username

Teams	Username
Aalto University	rdmaworkshop01
Azerbaijan State Oil and Indst. Uni.	rdmaworkshop02
Beihang University	rdmaworkshop03
CIUK	rdmaworkshop04
Indian Institute of Technology Kanpur	rdmaworkshop05
Nanyang Technological University	rdmaworkshop06
National Center for HPC, TAINCHC	rdmaworkshop07
National Center for HPC, LISΛRDGΓY	rdmaworkshop08
National Tsing Hua University	rdmaworkshop09
National University of Defense Technology	rdmaworkshop10
National Yang Ming Chiao Tung University	rdmaworkshop11
Peking University	rdmaworkshop12
ShanghaiTech University	rdmaworkshop13
Shanxi University	rdmaworkshop14
Southern University of Science and T.	rdmaworkshop15
Stony Brook University	rdmaworkshop16
Texas A&M University	rdmaworkshop17
University of Helsinki	rdmaworkshop18
University of Luxembourg	rdmaworkshop19
University of Maribor	rdmaworkshop20
Zhejiang University	rdmaworkshop21

Getting Started with HPC-AI AC Clusters

Once you got your username, login to the clusters.

To check available iris nodes using slurm commands.

$ sinfo -p iris
PARTITION AVAIL  TIMELIMIT  NODES  STATE NODELIST
iris         up   infinite     32   idle iris[001-032]

To allocate nodes interactively

Please avoid allocating nodes interactively if possible or set the time limit short because we are sharing the resources with multiple users.

$ salloc -N 2 -p iris -t 1:00:00
$ squeue -u $USER
     JOBID  PARTITION  NAME     USER      ST    TIME  NODES NODELIST(REASON)
     591608      iris interact <username>  R     0:24     2 iris[001-002]

To submit a batch job

$ sbatch -N 4 -p iris --time=1:00:00 <slurm script>

Please note that Iris has maximum 4 hour time limit.

Storage

We have limited disk space on nfs home. Please run jobs from the Lustre partition for your work,

When you login you are in the $HOME. There is also extra scratch space.
nfs home -> /global/home/users/$USER/
Lustre -> /global/exafs/users/$USER/

Please do not overload the cluster by submitting too many jobs.

If you have any issue, contact David Cho on Slack.

Getting Started with Iris Cluster - ISC26 SCC