Gaussian Tutorial
Command Line Basics
Command
Description
ls
list directories & files in current directory
mkdir newDir
makes a new directory in the current directory
cd ..
back up one directory
cd myDirectory/
open myDirectory
nano file1
opens file1 in nano editor
Ctrl-o , Enter
In nano-saves edits
Ctrl-X
Exits out of nano
Getting Started
MobaXterm
MobaXterm can be freely downloaded here (use the Home Installer Edition): https://mobaxterm.mobatek.net/download.html
Open MobaX.
Log in to Lawrence (refer to the following code).
[user.name@ ~]$ ssh User.Name@Lawrence.usd.edu
Warning: Permanently added 'lawrence.usd.edu' (RSA) to the list of known hosts.
User.Name@Lawrence.usd.edu's password:
Last login: Wed Jan 9 13:14:19 2019 from 192.236.35.187
Welcome to the Lawrence Supercomputer at the University of South Dakota!
......
[user.name@usd.local@login ~]$Transferring Files (Lawrence to & from PC)
In MobaX, the file explorer/hierarchy is visible just to the left of the command line interface. Navigate into directories by clicking on them, and out of directories using the "parent directory" button.
Note: entering/exiting a directory in this area doesn't change your current directory in the command line (you'll still be in the same directory as before in the command line). Use the cd command to navigate directories in the command line.
From this section of the MobaX window, you can drag and drop files between Lawrence and your PC as needed.
Once your files are in the proper directories, they can be handled in the same way as the example input files are handled in this tutorial (excluding the steps of copying the example input files into their respective directories).
Interactive Jobs
Open MobaX and log into Lawrence (see section "Getting Started").
Make a directory in your home directory, and navigate into it:
[user.name@usd.local@login ~]$ mkdir myInteractiveDir
[user.name@usd.local@login ~]$ cd myInteractiveDir/
[user.name@usd.local@login myInteractiveDir]$Connect to a node.
[user.name@usd.local@login myInteractiveDir]$ srun --pty bash
[user.name@usd.local@node56 myInteractiveDir]$Copy the Gaussian test file into your myInteractiveDir directory.
[user.name@usd.local@node56 myInteractiveDir]$ cp /opt/examples/Gaussian/test.com ./test.com
[user.name@usd.local@node56 myInteractiveDir]$ ls
test.comTo see the Gaussian modules available:
[user.name@usd.local@node56 myInteractiveDir]$ module avail gauss
---------------------------- /act/modulefiles --------------------------------
gaussian/09 gaussian/16Load one of the Gaussian modules.
[user.name@usd.local@node56 myInteractiveDir]$ module load gaussian/16
[user.name@usd.local@node56 myInteractiveDir]$ module list
Currently Loaded Modulefiles:
1) gaussian/16Run Gaussian with an input file:
[user.name@usd.local@node56 myInteractiveDir]$ g16 <test.com >interactiveOutput.com
[user.name@usd.local@node56 myInteractiveDir]$To view the output, open "interactiveOutput.com" with nano or another editor.
[user.name@usd.local@node56 myInteractiveDir]$ nano interactiveOutput.comNavigate out of "myInteractiveDir" into your home directory:
[user.name@usd.local@node56 myInteractiveDir]$ cd ..
[user.name@usd.local@node56 ~]$ Exit the node.
[user.name@usd.local@node56 ~]$ exit
[user.name@usd.local@login ~]$ Batch Jobs
If entering the tutorial at this point, open MobaX and log into Lawrence (see the second half of "Getting Started").
Make a directory in your home directory, and navigate into it:
[user.name@usd.local@login ~]$ mkdir myBatchDir
[user.name@usd.local@login ~]$ cd myBatchDir
[user.name@usd.local@login myBatchDir]$Copy the Gaussian batch job template files into your myBatchDir directory.
[user.name@usd.local@login myBatchDir]$ cp -r /opt/examples/Gaussian/. ./
[user.name@usd.local@login myBatchDir]$ ls
gaussianBatchTemplate.sh output.txt test.comOpen the file "gaussianBatchTemplate.sh" using nano (or another editor, if you prefer) and note the features described in the image below.
[user.name@usd.local@login myBatchDir]$ nano gaussianBatchTemplate.shOpen the file "test.com" (you will have to exit viewing gaussianBatchTemplate.sh in nano to do this in the same terminal).
[user.name@usd.local@login myBatchDir]$ nano test.comIn the last line, "test.com" is the input file name and the slurm ID followed by ".out" is the output file name. Also note that the number of CPUs (ntasks) requested matches the %nprocshared (or %nprocs) in the input file.
--ntasks in batch file should equal %nprocshared in input file
Exit nano (Ctrl-x), then run the batch job:
[user.name@usd.local@login myBatchDir]$ sbatch gaussianBatchTemplate.sh
Submitted batch job 12345
[user.name@usd.local@login myBatchDir]$ Show the list of jobs running on Lawrence:
[user.name@usd.local@login myBatchDir]$ squeue
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
12345 nodes gaussian user2 R 0:22 1 node40
48469 nodes bash user1 R 3:04 1 node56
48467 nodes bash user3 R 23:07 5 node[25-29]
48458 nodes bash user4 R 23:41 1 node21
48459 nodes bash user4 R 23:41 1 node22
48460 nodes bash user4 R 23:41 1 node24
48457 nodes bash user4 R 25:14 1 node11
48449 nodes bash user4 R 25:18 1 node13
48450 nodes bash user4 R 25:18 1 node17
Show only your jobs:
[user.name@usd.local@login myBatchDir]$ squeue -u user.name
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
12345 nodes gaussian user.n R 2:34:52 1 node43When your job is complete, it will not appear on squeue. To see the output of your job when completed, open the ".out" file.
[user.name@usd.local@login myBatchDir]$ nano 12345.outTo see how the job went, and look at any errors, open the slurm file.
[user.name@usd.local@login myBatchDir]$ nano slurm-12345.out
Graphical User Interface Jobs
If entering the tutorial at this point, open MobaX and log into Lawrence (for help, see the section "Getting Started").
Starting the GUI
Log into a node
[user.name@usd.local@login ~]$ srun --pty bash
[user.name@usd.local@node51 ~]$ Load the Gaussian module:
[user.name@usd.local@node51 ~]$ module load gaussian/16
[user.name@usd.local@node51 ~]$ module list
Currently Loaded Modulefiles:
1) gaussian/16Launch the Gaussian GUI:
[user.name@usd.local@node51 ~]$ srun gviewThe GUI will open:
Last updated