This covers chemically dissimilar films where one film is one of the 16 original groups, against the 3 new candidates.
Note: I did this in the /ccs/proj/ directory on Titan
Note: Titan has been decomissioned as of August 01, 2019; some of these commands might not work on other systems without editing.
>> conda create --name myconda python=3.5
>> source activate myconda
Note: You may have to first append to your path the directory where anaconda is located, e.g.
>> export PATH=/ccs/proj/xxx000/anaconda/titan/bin:$PATH
commit caa89f99a07c84be4c457378138105100eaa9dfb
>> git clone https://github.com/PTC-CMC/atools.git
>> pip install .
commit fa2bc651823d8c0a93cac8721e0abf10a7b5e168
>> git clone https://github.com/mosdef-hub/mbuild.git
>> pip install .
commit 1aa97bbebed22c94ad8d9d68486fbdbe7a3bd6d7
>> git clone https://github.com/mosdef-hub/foyer.git
>> pip install .
>> conda install signac-flow=0.5.4 -c glotzer
Note: If mBuild and Foyer are installed via conda or pip, these dependencies should be installed automatically.
>> conda config --add channels omnia mosdef
>> conda install lxml requests networkx mdtraj oset parmed openmm plyplus
>> pip install mdanalysis
>> git clone https://github.com/PTC-CMC/terminal_group_mixed_original_16_new_3.git
Note: All flow commands must be performed from the project root directory.
Note: The -n 3 -c 17 1
signifies that three statepoints will be created for each
parameter state, with an alkane backbone of 17 carbons, each with a different random seed (incrementing from 1)
>> python src/init.py -n 3 -c 17 1
This will submit jobs in bundles of 6 statepoints to be executed on a single node. Although each node contains 16 processors, memory issues limit the number of simultaneous systems that can be initialized.
>> python src/project.py submit -o initialize_system --bundle 6 --nn 1 -w 0.5
>> python src/project.py submit -o fix_overlaps --bundle 400 --nn 400 -w 1
>> python src/project.py submit -o lmp_to_gmx --bundle 48 --nn 3 -w 0.5
>> python src/project.py submit -o em_grompp --bundle 400 --nn 400 -w 0.5
>> python src/project.py submit -o mdrun_em --bundle 400 --nn 400 -w 0.5
>> python src/project.py submit -o nvt_equil_grompp --bundle 400 --nn 400 -w 0.5
>> python src/project.py submit -o nvt_mdrun --bundle 400 --nn 400 -w 2
>> python src/project.py submit -o compress_grompp --bundle 400 --nn 400 -w 0.5
>> python src/project.py submit -o mdrun_compress --bundle 400 --nn 400 -w 1
>> python src/project.py submit -o shear_5nN_grompp --bundle 400 --nn 400 -w 0.5
Note: Shear was originally performed for 5ns and then extended another 5ns. The MDP files have been updated to include the full 10ns now.
>> python src/project.py submit -o mdrun_shear_5nN --bundle 400 --nn 400 -w 4
>> python src/project.py submit -o shear_15nN_grompp --bundle 400 --nn 400 -w 0.5
>> python src/project.py submit -o mdrun_shear_15nN --bundle 400 --nn 400 -w 4
>> python src/project.py submit -o shear_25nN_grompp --bundle 400 --nn 400 -w 0.5
>> python src/project.py submit -o mdrun_shear_25nN --bundle 400 --nn 400 -w 4
NOTE: If the simulations do not complete in the hours provided, concatenating the TRR files is necessary.
This is easier to do through the trjcat
command from command line GROMACS
.
gmx_mpi trjcat -f shear_5nN*.trr -o shear_5nN_combined.trr
gmx_mpi trjcat -f shear_15nN*.trr -o shear_15nN_combined.trr
gmx_mpi trjcat -f shear_25nN*.trr -o shear_25nN_combined.trr
>> python src/analysis.py submit -o calc_friction_system --bundle 48 --nn 3 -w 1
>> python src/analysis.py submit -o calc_cof --bundle 48 --nn 3 -w 1