Skip to content

Quick examples for running jobs on TACC in serial or parallel. Examples are for Frontera.

Notifications You must be signed in to change notification settings

michaelsegzy/TACC_Tutorial

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TACC_Tutorial

Quick examples for running jobs on TACC in serial or parallel. Examples are for Frontera, but are nearly identical for all clusters.

1. Login via ssh

  • On Mac open Terminal. Application > Utilities > Terminal
  • You need to choose TFA option in your TACC user portal. I find downloading the app faster and more secure than using text messaging.
ssh example_user@frontera.tacc.utexas.edu
enter password
enter code for 2-factor authentication

2. Clone/fork this repository or make one

  • Once connected to TACC you are on a login node. Login nodes are a shared resource with other users. Do NOT run or do anything that requires computation/resources. You can copy things to Corral (long-term storage), but other jobs must be done on a compute node.
  • cdw = change directory to $WORK, same as cd $WORK. TACC has 3 main directory $HOME, $WORK, $SCRATCH
  • You will need to become familiar with Linux and BASH to work efficiently in a high performance computing (HPC) environment
git clone https://github.com/emjavan/TACC_Tutorial.git
  • Example of initializing a git repo and pushing
mkdir practice_repo
cd practice_repo
git init
time for iter in {1..10}; do echo $iter; done >> test_loop.txt 
git add test_loop.txt
git commit -m "first commit, created test_loop.txt"
git push

3. Pick an editor

  • vi, vim, emacs. You'll need to become comfortable with some editor to make changes to files and debug

4. Request compute node

  • To test out some code quickly you'll want a development node, the max time is 2hrs. Replace Example_Allocation with the name of the allocation you've been added to.
idev -N 1 -n 1 -t 02:00:00 -p development -A Example_Allocation
  • If you need more time and don't want to submit a job you will have to wait in the queue live and risk your internet connection being broken and disrupting your work.
idev -N 1 -n 1 -t 6:00:00 -p small -A Example_Allocation
  • If you have a very large job you may wish to avoid multiple tasks being performed on one node and/or need a high memory node. nvdimm is more expensive than normal and small queues, so do not use until testing the limits of your code.
idev -N 1 -ntasks-per-node 1 -t 02:00:00 -p nvdimm -A Example_Allocation

4. Submit a serial job

  • Submit your job
sbatch my_serial_job.sh
  • Check on your job online at the TACC portal or showq = show queue
showq -u

5. Submit a parallel job

sbatch launch_parallel_job.sh

About

Quick examples for running jobs on TACC in serial or parallel. Examples are for Frontera.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Shell 100.0%