- Open the local directory in terminal.
- Open an SSH connection to Cheaha:
- Change directory to the desired directory:
cd brainPH
- Create a virtual environment to one directory up. So, we can pull the current directory to local machine again without the venv files.
python3 -m venv ../venvs/brainph
- Activate the environment:
source ../venvs/brainph/bin/activate
- Upgrade pip:
pip install --upgrade pip
- Install the requirements:
pip install -r requirements_cheaha.txt
- Create a batch job, for example: randomdistance.job
- Run the Job in Cheaha:
sbatch randomdistance.job
sbatch randomcluster.job
- To kill a Slurm job
scancel <jobid>
You can find your jobid with the following command:
squeue -u $USER
- See the users in a partition:
squeue -p amd-hdr100
- Check details of a job, (13626634 is the job id):
squeue -j 13626634 -o "%all"
- See the running scripts for the user:
ps aux | grep ashovon
- Get a medium instance:
srun --ntasks=1 --cpus-per-task=6 --mem-per-cpu=8192 --time=50:00:00 --partition=medium --job-name=fmri --pty /bin/bash
srun --ntasks=4 --cpus-per-task=6 --mem-per-cpu=8192 --time=50:00:00 --partition=medium --job-name=fmri_1400 --pty /bin/bash
- Download a file from Cheaha to local:
scp [email protected]:/home/ashovon/brainPHmatlab/fcn_645.png .
scp [email protected]:/home/ashovon/brainPH/randomcluster_2_job.out .- Download all files from Cheaha directory to local:
scp -r [email protected]:/home/ashovon/brainPH/random_data_1 .
scp -r [email protected]:/home/ashovon/brainPHmatlab/full_data_linear .- Upload a file from local directory to Cheaha
scp timeseries.Yeo2011.mm316.mat [email protected]:/home/ashovon/brainPHmatlab/