My wiki | Site / RunningGibbsHelmholtz

Hemholtz and Gibbs are the Lobster Room machines, which can run a limited number of jobs. Priority is given to the Beveridge/Thayer lab, but is an alternative place to run Amber18.

Submitting Jobs

In order to submit jobs to helmholtz/gibbs you will need to ssh in:

   ssh username@helmholtz.computer.wesleyan.edu

All jobs are submitted directly to the machine instead of to a scheduler. This means that if you submit a job to a GPU that is already in use, you will terminate someone's job! Use:

 
   nvidia-smi

This will give an output table with Volatile GPU-Util. This number will indicate whether or not a GPU is in use. Use nohup to submit a job to an open GPU so that it will continue to run when you close the terminal:

   nohup sh jobscript.sh

helmholtz/gibbs CPU Header

# cd to working directory
cd /data/Kristen/CPU

# call Amber18
sander.MPI -O -i inputs.in -p prmtop_wat.prmtop -c coordinates_in.rst -r coordinates_out.rst -ref restraint_reference.rst -o energy_info.out

helmholtz/gibbs GPU Header

# specify which GPU to use, where n=0,1,2, or 3
export CUDA_VISIBLE_DEVICES=n

# cd to working directory
cd /data/Kristen/GPU

# call Amber18
mpirun -np 1 $AMBERHOME/bin/pmemd.cuda.MPI -O -i inputs.in -p prmtop_wat.prmtop -c coordinates_in.rst -r coordinates_out.rst -ref restraint_reference.rst -o energy_info.out -x trajectory.mdcrd