Gaussian03
To run Gaussian 03 on any of our clusters:
1. type: module load gaussian
2. Edit your gaussian input script file (filename.com) to put "%NProc=n" on the top line, where n may be 2 or 4 or 8. Gaussian will create no more than n threads during the job. Your nproc shouldn't exceed the number of cpus you'll reserve in your PBS script file. On your next line put "%mem=nmb" which tells Gaussian how much memory to allocate. Typical amounts are 1800 or 3800. Nyx has nodes with 2,4,6,8 or 64GB. Requesting a maximum allocation of 1800mb leaves room on a 2gb node for the OS, the application,etc.
example of top lines of input file
%nproc=4
%mem=3600mb
3. Write an appropriate PBS script file for your job. Set the queue to route if you need no more than 7600mb; otherwise set it to bigmem, e.g.
#PBS -q route
or
#PBS -q bigmem
Next set the nodes resource to the minimum number of CPUs you'll need. Remember that all the nodes on nyx have at least two CPUs each. Since gaussian is an SMP code you'll need all the CPUs required on one node. So on nyx your choices are:
#PBS -l nodes=1:ppn=2
or
#PBS -l nodes=1:ppn=4
or
#PBS -l nodes=1:ppn=8
remember to get eight CPUs you must use the bigmem queue.
if you need more than 3600mb you should set the pmem resource as well, e.g.
#PBS -l pmem=6600mb
example of a PBS script file for gaussian:
#PBS -N g03_test
#PBS -q route
#PBS -l nodes=1:ppn=4,pmem=5600mb,walltime=24:00:00
#PBS -S /bin/csh
#PBS -M your-email-address
#PBS -m abe
#PBS -V
#
echo "I ran on:"
cat $PBS_NODEFILE
#
#cd to your execution directory first
cd /home/your-user-name
#
#
g03 < gaussian.input > gaussian.output
another example of a PBS script file for gaussian:
#PBS -N g03_bigtest
#PBS -q bigmem
#PBS -l nodes=1:ppn=8,walltime=24:00:00
#PBS -S /bin/csh
#PBS -M your-email-address
#PBS -m abe
#PBS -V
#
echo "I ran on:"
cat $PBS_NODEFILE
#
#cd to the directory you were in when you submitted the job
cd $PBS_O_WORKDIR
#
#
g03 < gaussian.input > gaussian.output
4. submit the job with: qsub name_of_script_file