Gaussian predicts the energies, molecular structures, and vibrational frequencies of molecular systems, along with numerous molecular properties derived from these basic computation types. It can be used to study molecules and reactions under a wide range of conditions, including both stable species and compounds which are difficult or impossible to observe experimentally such as short-lived intermediates and transition structures. We have Gaussian g03 D.01 with Linda 7.1 installed.
To load Gaussian type the following:
$ module load md/gaussian-g03.d01
Gaussian and Linda are both commercial software. They require licenses to use. If you have a license let us know (send an email to help@cse) and we will enable access for you to use this software.
Here is an example submit script that will run test #727 from the distribution using 8 cpus. To submit just type qsub submit.sh
.
#!/bin/sh #$ -cwd #$ -j y #$ -S /bin/bash #$ -N gausstest #$ -e 727.err #$ -o 727.out #$ -pe linda 8 # Load the gaussian environment module load md/gaussian-g03.d01 infile=~/gaussian/727/test727.com tmpinfile=$TMPDIR/infile.com outfile=~/gaussian/727/test727.log # Populate the input file with the nodes SGE gives us # nprocshared = number of cpus per node # nproclinda = number of nodes # lindaworkers = CSV hostnames (ie "ecompute-5-1,ecompute-5-2") awk -v nprocshared="4" -v nproclinda="2" -v lindaworkers="`cat $TMPDIR/machines`" -f $g03root/sge_linda.awk $infile > $tmpinfile time g03l < $tmpinfile > $outfile
The awk line in the above script just makes sure Gaussian knows how many CPUs you have been allocated, how many nodes, and which nodes. The only thing that you need to modify is the nproclinda option.
Gaussian allows you to specify where to put temporary files. This is specified in the input file. The following will use a file called scott.rwf in the directory /state/partition1 for scratch space.
%RWF=/state/partition1/scott %NoSave
The No Save
option removes the temporary file after the run completes.
For more information see this online documentation.
By default Gaussian 03 will put the checkpoint file into your current working directory. However you may want to save this file in order to restart your job or visualize your result. You can name your checkpoint file in the input file, for example by:
%Chk=/home/foo/gaussian/727/727
the checkpoint file will be saved in /home/foo/gaussian/727 as 727.chk.
Note: Saving the checkpoint file over NFS will likely slow your job down.
There is a help program installed called ghelp
. Once you load Gaussian into your environment it will be in your path. You can also browse the online documentation.
$ ghelp to get help on a particular topic type: ghelp <topic> or ghelp <topic> <sub-topic> ... '<topic>' or <sub-topic> may be the name of some particular information desired, one of the special words 'topics' or 'options', or the wild card symbol '*'. abbreviations are allowed for any name, but no checking for ambiguity is done. example: -------- ghelp route rhf Other information available: buckets common fileio geometry ghelp introduction l101 l103 l202 l401 ov4 l502 l503 l506 l601 l602 l1002 l9999 loop ov1 ov3 ov5 ov6 ov7 ov8 ov9 putil recursion recursion restart route links rwf references util general_basis Utilities c8695 chkmove cubman cubegen g03opt ham506 mm formchk newzmat rwfdump testrt unfchk Browse Archive arctmp crendx reform