Skip to Content

High Performance Computing


Here is some documentation which may help you to run HADDOCK 2.1 on Flux.  Files from the file server are also mounted there, to ease copying and moving.

Step 1: Login to flux and load the modules for haddock:

$ ssh -X
$ module load lsa
$ module load cns/1.2
$ module load haddock

Step 2: Move your files from the department file server to the scratch space on flux:

$ cd /dept/brd/[home/group]/[username/groupname]

$ cd /dept/brd/home/bjensen
$ cp -r p450 /nobackup/bjensen/

Step 3: Move to the scratch space, and adjust the new.html file to reflect the path to your current directory. Example given below.  Items to be changed are in green.

$ cd /nobackup/bjensen/p450
$ nedit new.html

<title>HADDOCK - start</title>
<body bgcolor=#ffffff>
<h2>Parameters for the start:</h2>
<h4><!-- HADDOCK -->
UNAMBIG_TBL=/nobackup/bjensen/p450/pre_p450_b5_ NEWcuriesol_U13C15N_rem32333435_02282011.tbl<BR>
submit_save=Save updated parameters<BR>
</h4><!-- HADDOCK -->

Now run haddock to setup the run:
$ haddock

Step 4: Modify anything in the run folder that needs to be changed.  Examples:

  • files in protocols
  • files in toppar
  • others

Step 5: Copy the haddock tools to the run directory:

$ cd /nobackup/bjensen/p450/run3/
$ cp /home/software/rhel5/lsa/haddock2.1/haddock_setup.tgz .
$ tar -xvf haddock_setup.tgz

Step 6: Modify the following line in the haddock.pbs file to reflect the amount of time it will take for the entire caluclation to take. Change walltime to reflect hh:mm:ss. Guidelines:

  • 500: ~6-8 hours
  • 1000: ~14-18 hours
  • 2000: ~36-48 hours
#PBS -N haddock
#PBS -l nodes=1:ppn=1, walltime=16:00:00,qos=biophysics_flux
#PBS -m abe
#PBS -q flux
#PBS -j oe
#PBS -A biophysics_flux
#PBS -o haddock.out
#PBS -e haddock.err

Step 7: Modify the run.cns to setup your experiment also make sure to include the proper paths to the run (e.g. /nobackup/bjense/p450/run3).

Step 8: Execute the job:

$ qsub haddock.pbs
$ qstat -uuniqname
to check the status