Useful Links:

https://wiki rc.usf.edu

https://wiki.rc.usf.edu/index.php/Connecting_To_CIRCE

https://wiki.rc.usf.edu/index.php/Guide_to_SLURM

https://wiki.rc.usf.edu/index.php/CIRCE_Hardware

Some Instructions:

************************ Run via X2Go
run x2goclient
enter net ID and password

ssh mdc-1057-13-8

module load apps/cuda/12.2.2
deviceQuery

Download anaconda with wget and install it
create your virtual env to work in

When the ssh connection is lost, the process terminates. If you do not want that to happen you need to use SLURM to submit you jobs.

For Slurm
You can submit jobs to the new GPU node with SLURM using the muma_2021 partition and muma21 as the qos. To get your job assigned GPUs on the new node, you will need to add the following flag

--gres=gpu:1

************************ useful commands *************************************

cd ~ : your home folder

cd /: root (everyones account is there)

cd /work_bgfs/user (your working directory)

mv mv ~/coach-master/ /work_bgfs/user : move files to work directory (work_bgfs)

sinfo : list compute nodes

du -hc --max-depth=1 --inodes : quota of inondes in home: max is about 400k:

************************ File transfer *************************************
sftp username@circe.rc.usf.edu


put filename

lpwd: local working directory
pwd: remote working directory
lls: local
ls:remote

************************ Slurm script *************************************
#!/bin/bash
#SBATCH --time=6:00:00
#SBATCH --partition=muma_2021
#SBATCH --qos=muma21
#SBATCH --gres=gpu:1
#SBATCH --output=/home/e/use/Documents/output.%j
#SBATCH -e error.%j.%N.txt

module load apps/cuda/12.2.2

python your_python_code.py


Interactive Session:

srun -p muma_2021 -q muma21 --gres=gpu:1 -t 06:00:00 --pty /bin/bash
srun -p henderson_itn18 -q hen18 --gres=gpu:1 --pty /bin/bash
srun -p henderson_itn18 -q hen18 --gres=gpu:1 --mem=6G -N 1 -n 4 -t 9:55:00 --pty /bin/bash

srun --partition=cbcs --qos=preempt --nodes=1 --ntasks=1 --cpus-per-task=6 --time=65:00:00 --mem=120gb --gres=gpu:1 --pty /bin/bash

srun --partition=chbme_2018 --qos=preempt --nodes=1 --ntasks=1 --cpus-per-task=6 --time=110:00:00 --mem=5gb --gres=gpu:1 --pty /bin/bash


************************ Instructions on installing Miniconda3 *************************************
    cd $HOME
    wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
    chmod +x Miniconda3-latest-Linux-x86_64.sh
    ./Miniconda3-latest-Linux-x86_64.sh
        when the installer asks “Do you wish the installer to initialize Miniconda3 by running conda init?”, please answer "yes"
    log-out of CIRCE/RRA, then log back in
    which conda pip
        make sure it returns with: ~/miniconda3/bin/conda and ~/miniconda3/bin/pip
    pip install --upgrade pip
    conda update --all -y
    conda install -y anaconda

Once Conda is installed, you can then create virtual environments for each package (or combination of packages) that you may need, including Tensorflow, Keras, OpenCV, gurobipy, numpy, etc.

A step-by-step walkthrough of using Conda (including creating Virtual Environments) is located at: Getting started with conda — conda 22.9.0.post96+2abc88b82 documentation  

************************ help email *************************************

rc-help@usf.edu