Using Gaussian - mghpcc-projects/rc_howtos GitHub Wiki
#Getting Started
The following guides were created based on the tutorial found here:
www.uio.no/english/services/it/research/hpc/abel/help/software/Gaussian.html
It is strongly reccomended to make a new directory before running tests
mkdir /home/$USER/gaussianwork && cd /home/$USER/gaussianwork
The examples in this guide will use the hexclethane.com file. You can either:
-
Right click the 'hexclethane.com' link below and click 'Save Link As'
-
Copy and paste the information into a blank file and name it hexclethane.com
click here to download hexclethane.com
# HF/MP2 6-31g opt Ethane 0 1 C1 C2 C1 RCC Cl3 C1 RCH C2 ACCH Cl4 C1 RCH C2 ACCH Cl3 T1 Cl5 C1 RCH C2 ACCH Cl3 T2 Cl6 C2 RCH C1 ACCH Cl3 T3 Cl7 C2 RCH C1 ACCH Cl6 T4 Cl8 C2 RCH C1 ACCH Cl6 T5 RCH = 1.6 RCC = 1.2 ACCH = 115.2 T1 = 125.0 T2 = -125.0 T3 = 185.0 T4 = 125.0 T5 = -125.0
#Running Gaussian in an interactive session
To run Gaussian, you will need to start up an interactive session via SLURM
srun --pty -N 1 --exclusive -p defq /bin/bash
Load the Gaussian Module for c3ddb You will need to add the gaussian module each time you login to c3ddb
module add c3ddb/gaussian
The variable PGI_TERM needs to be unset or else it will cause script errors
unset PGI_TERM
The example in this guide will use the hexclethane.com file shown above.
Run your .com job file and produce a log file
g09 < hexclethane.com | tee /home/$USER/gaussianwork/hexclethane.log
#Running Gaussian from a batch script
If you would rather run your own .com test then you will need to copy the .com file of your choice into your working directory, for this example we use the hexclethane.com file.
To automate this process you can use a simple script like the one shown below by either:
* Downloading the script by right clicking the 'gaussiantest.sh' link and clicking 'Save Link As' * Creating your own script by copying the script below and adjusting the relevant data.click here to download gaussiantest.sh
#!/bin/bash #SBATCH --nodes=1 #SBATCH --ntasks-per-node=8 #SBATCH --job-name=myHexclethane #SBATCH --time=0:50:0 #SBATCH --mem-per-cpu=1G #SBATCH --partition defq module add c3ddb/gaussian unset PGI_TERM g09 < hexclethane.com | tee /home/$USER/gaussianwork/hexclethane.log
This simple script when filled out with the appropriate user data will run your com file and place a copy of the log file into a directory of your choice simply by using the sbatch command followed by your scripts name.
>sbatch gaussiantest.sh##Here is a good reference for running Gaussian tests: