DSAI Quick Start
What you need
Item |
Notes |
|---|---|
JHED and DSAI account |
Request one on Coldfront |
Hopkins VPN (Pulse Secure) |
Required from off-campus for some services |
SSH client |
macOS/Linux: built-in • Windows: OpenSSH or PuTTY |
Log in
ssh <YourUserID>@dsailogin.arch.jhu.edu
First time? Type yes at the host-key prompt, then your DSAI password.
Submit your first batch job
Create hello.slurm :
#!/bin/bash
#SBATCH --job-name=hello
#SBATCH --output=hello.%j.out
#SBATCH --time=00:02:00
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=1
#SBATCH --partition=$Partition
echo "Hello from $(hostname)!"
sleep 30
Submit & monitor:
sbatch hello.slurm # submit
squeue -u $USER # check status
tail -f hello.<jobID>.out
If the job is PENDING for more than a minute, reduce resources or pick another partition (see table below).
Partitions at a glance
For available partitions, see: Available Partitions
Managing Python, Jupyter, RStudio
Work in Progress
Storage locations
Path |
Default quota |
Intended for |
|---|---|---|
|
50 GB (backed up) |
configs, notebooks, small scripts |
|
1 TB |
working data |
For more information on available filesystems, see here: Filesystems on DSAI
House-keeping
Fair-share scheduler – large jobs may wait if your lab has used more CPU-hours than average recently.
Login nodes – no heavy compute. Use
sbatchinstead.
Need help?
Email: help@arch.jhu.edu