-
Notifications
You must be signed in to change notification settings - Fork 0
Initial Regular MD
We use a SLURM cycler script to efficiently schedule our production runs.
This is a suggested file that cycles the Molecular Dynamics simulations for n times.
Right now, this is hard-coded for the University of Rochester Bluehive Slurm shared environment but can be used in any Slurm environment given certain parameters are changed such as partition name, module name etc.
This is not a difficult task if you already have a Slurm job script that you have used for GROMACS runs previously, if not carefully go through the script to see what it is doing.
Once again, this is a suggested practise not critical for the pipeline.
If you don't have to cycle the MD simulations (if you are working in a private cluster or does not have a time-limit for a Slurm job) or have an infrastructure that's not Slurm-like, feel free to use your own script to run the GROMACS engine.
You can find the cycler-related scripts in the directory
Templates/Cycler/
. Go through the test
directory for Cycler carefully before using it.
An example scenario to test this cycler can be found there.
Also, note that the cycler demands renaming of certain files (.cpt
and .gro
files from the last equilibration step (6.6), and .mdp
file) in a specific way.
.mdp
file for production run should be modified such that it is compatible with the cycler. Also, here we need separate .mdp
files for
323K, and 423K. Sample .mdp
files can be found
Templates/Production/
.
Separating pre- and post-production files
into different folders is highly recommended.
A script can be used to
automate this task to be executed from
/PATH/TO/YOUR/UNZIPPED/FOLDER/gromacs/
. Given below is a script that can do
it for you. This is convenient if you have multiple replicates and systems.
CAUTION! This is a suggested template and your file tree might be different and therefore might do irreversible things.
Go through this script and understand what it does to your files before executing it. Note that the given script here is for the DIPC system and has to be replaced accordingly for the DIPC system.
copy_script.sh
#############################################
# Usage: #
# bash copy_script.sh #REPLICANUMBER# #
#############################################
CURRENT=$PWD
for T in 323K 423K
do
dest=../../Initial-Regular-MD/0${1}/$T
mkdir -p $dest/
cp ../../Templates/Production/${T}/step7_production.mdp $dest/
cp step7_production.cpt $dest/
cp step7_production.gro $dest/
cp system.top $dest/
cp index.ndx $dest/
cp -r toppar/ $dest/
cp ref_system_centered.pdb $dest/
cd $dest/
mv step7_production.cpt $1"DIPC_0.cpt"
mv step7_production.mdp $1"DIPC.mdp"
mv step7_production.gro $1"DIPC.gro"
cp ../../Templates/Cycler/cycle* .
cd $CURRENT
done
Example usage for replica 1 : (To be executed from your simulation dir for replica 1)
bash copy_script.sh 1
This would result in 2 new dir named EM_Equilibration and Production in your
sim dir. All the files necessary for starting the production run for each Temperature are ready inside Initial-Regular-MD/
dir. In order to run the cycler and
therefore to start the production NPT run, cd
into the respective
temperature dir inside the Initial-Regular-MD/
dir and execute the following:
./cycle 1DIPC-1