Vista Quickstart: Access and Submitting a Job on Vista

In this quickstart, we will walk through the steps to access the Vista compute system from TACC and submit a simple job. This quickstart assumes you are eligible for or affiliated with projects that have allocations on Vista. While Linux experience is helpful, it is not required to complete the quickstart.

Before you can access the Vista computing system, you must:

  1. Have a TACC account or register for a TACC account
  2. Be on an active project with allocation for Vista or submit an allocation request
  3. Set up Multi-Factor Authentication at TACC

  1. Open a shell, such as Terminal on macOS/Linux or PowerShell on Microsoft Windows. Terminal screenshot
  2. In the shell, enter the following: SSH command screenshot
  3. When prompted, enter your TACC account password.
  4. When prompted, open your Multi-Factor Authentication application and enter the current token displayed. Authentication screenshot
  5. If authentication is successful, you will be logged into a Vista login node First login screenshot

  1. If you are not familiar with Linux commands, we recommend starting with the Introduction to Linux roadmap.
  2. List the files in the current directory by typing ls. If this is your first time logging in, nothing is returned. ls command screenshot
  3. Print the working directory by typing pwd. pwd command screenshot
  4. Create a new subdirectory, vista_quick_start, with the mkdir command: mkdir vista_quick_start. mkdir command screenshot
  5. Navigate to vista_quick_start with the command cd: cd vista_quick_start. We will create files in this directory. cd command screenshot

User environment and software are managed by environment modules. You may read more about them at Modules on Vista. This brief section will walk through basic module usage.

  1. See the currently loaded modules by typing module list. module list screenshot
  2. See all available modules on Vista by typing module avail. Navigate the list of modules with up and down keys, and quit navigation by pressing the Q key. module avail screenshot
  3. List more details about a specific module by typing module spider <modulename> or module spider <modulename/version>. For example, to see the module gcc/15.1.0, type module spider gcc/15.1.0. module spider screenshot
  4. Load a module by typing module load <modulename>. For example, to compile a C program with the GCC of version 15.1.0, type module load gcc/15.1.0. module load screenshot

You will compile a simple C program in this section.

  1. Download squares.c into your current directory with wget: wget https://cvw.cac.cornell.edu/cintro/wrapping-up/squares.c wget command screenshot
  2. Compile the program with gcc by entering the following: gcc -o square squares.c gcc command screenshot

Vista uses Slurm to schedule jobs. This means that to submit a job, you must create a job script.

  1. Create a script file named quickstart.sh with the touch command: touch quickstart.sh touch command screenshot
  2. First, copy the content below into your clipboard.
  1. Use the command nano to edit quickstart.sh: nano quickstart.sh nano command screenshot
  2. Paste in the script paste screenshot
  3. Edit the file as necessary. For example, myproject should be replaced with your project name.
  4. When you are finished editing, press Control and X at the same time on your keyboard.
  5. When prompted if you would like to save, press Y. save screenshot
  6. When prompted for file name to write, press enter without additional input. file name screenshot
  7. Check if quickstart.sh is saved correctly by entering: cat quickstart.sh cat command screenshot
  8. Submit the job with the command: sbatch square.sh sbatch command screenshot

Since the script is very brief and the program is short, it should be finished immediately. However, if the queues are busy or your job is still running, you can check its status with squeue: squeue -j <job_id> squeue command screenshot

  1. Once you have confirmed that the job is completed, type ls to list the files. There should be two more files that follow the scheme of myjob.o<job_id> and myjob.e<job_id>. ls after job screenshot
  2. The first file, myjob.o<job_id>, contains any standard output produced by your job. output file screenshot
  3. The second file, myjob.e<job_id>, contains any error messages produced by your job. In this case, the file is empty because the program is simple. error file screenshot
©  |   Cornell University    |   Center for Advanced Computing    |   Copyright Statement    |   Access Statement
CVW material development is supported by NSF OAC awards 1854828, 2321040, 2323116 (UT Austin) and 2005506 (Indiana University)