Full documentation here.
Login at this link.
- Open a terminal on the head node by selecting Clusters/Alpine
- Start a VS Code interactive job Interactive Apps/vS Code-Server (Presets)
Directions to set up ssh are here
Command to log in using the created key
ssh -i ~/.ssh/id_rsa {user_name}@login-ci.rc.colorado.edu
Submit a script
sbatch job_script.sh
Cancel a job
scancel job_id
Start an interactive job
sinteractive --partition=amilan --time=00:10:00 --ntasks=1
To view all jobs
squeue
To view just your jobs
squeue --user {user_name}
Set up Globus. Note, need to set up and use the CU Boulder Research Computing ACCESS
.
rsync -avz --progress path/to/local/files {user_name}@login-ci.rc.colorado.edu:/home/{user_name}/path/to/alpine/location
To start conda
module load anaconda
conda init
By default conda will install in the home directory, but there is not enough room. Instead it needs to install into projects.
vim ~/.condarc
And add
pkgs_dirs:
- /projects/$USER/.conda_pkgs
envs_dirs:
- /projects/$USER/software/anaconda/envs
Now when you create a new environment, it will point to the correct location
Information for chainging the base directory of installing packages with conda is here
module load slurmtools
See usage stats
suuser {user_name}
See usage stats for the last year
suuser {user_name} 365
This is important to check because
The cumulative computing allocations of a single research group across all projects may not exceed 5 M SU/year or 2.5 M SU over 6 months
Check on jobs that have run (including wait times)
jobstats {user_name}
Check priority (greater than 1 is higher than average, less than 1 is lower than average)
levelfs {user_name}
View resources requested by a job
scontrol show job {job_id}