I try to run C-G MD using gromacs. For this purpose I used the martini ff and I follow martini tutprial http://md.chem.rug.nl/cgmartini/index.php/proteins#martini.
However on equilibration stage I am having strange error:
Step 37, time 0.74 (ps) LINCS WARNING
relative constraint deviation after LINCS:
rms 97.314256, max 5264.672363 (between atoms 10452 and 10454)
bonds that rotated more than 30 degrees:
atom 1 atom 2 angle previous, current, constraint length
10441 10443 90.5 0.3100 1.7984 0.3100
10443 10445 90.1 0.3101 64.6521 0.3100
10445 10447 90.0 0.3098 592.6801 0.3100
10447 10450 89.7 0.3112 839.5673 0.3100
10450 10452 90.0 0.3022 992.5125 0.3100
10452 10454 90.3 0.2972 1632.3584 0.3100
10454 10457 90.2 0.3761 1480.2501 0.3100
10457 10459 89.0 0.3836 45.4105 0.3100
10459 10463 90.9 0.3092 87.7713 0.3100
10463 10467 89.3 0.3084 10.4813 0.3100
10467 10469 92.2 0.3096 1.0514 0.3100
10469 10471 82.3 0.3139 0.2427 0.3100
10471 10473 90.1 0.3142 0.3754 0.3100
10473 10475 90.9 0.3110 0.6999 0.3100
10475 10477 90.0 0.3096 0.3452 0.3100
Wrote pdb files with previous and current coordinates
mpiexec: process_obit_event: evt 6 task 0 on wn657 stat 267.
mpiexec: wait_tasks: waiting for wn657 wn731 wn731.
mpiexec: kill_others_now: alarm went off, killing all other tasks.
mpiexec: kill_tasks: killing all tasks.
mpiexec: process_kill_event: evt 10 task 1 on wn657.
mpiexec: process_kill_event: evt 11 task 2 on wn731.
mpiexec: process_kill_event: evt 12 task 3 on wn731.
mpiexec: process_obit_event: evt 8 task 2 on wn731 stat 265.
mpiexec: process_obit_event: evt 9 task 3 on wn731 stat 265.
mpiexec: process_obit_event: evt 7 task 1 on wn657 stat 265.
mpiexec: Warning: task 0 died with signal 11 (Segmentation fault).
mpiexec: Warning: tasks 1-3 died with signal 9 (Killed).
Max number of connections per atom is 20
Total number of connections is 154116
Max number of graph edges per atom is 7
Total number of graph edges is 44472
Constraining the starting coordinates (step 0)
Constraining the coordinates at t0-dt (step 0)
RMS relative constraint deviation after constraining: 4.04e-06
Initial temperature: 1.8161e-07 K
Started mdrun on node 0 Thu Jul 31 15:19:14 2014
Step Time Lambda
0 0.00000 0.00000
Grid: 42 x 42 x 42 cells
Energies (kJ/mol)
Bond G96Angle Proper Dih. Improper Dih. LJ (SR)
6.24872e+03 1.47844e+04 1.75778e+03 4.11810e+02 -2.32179e+07
Coulomb (SR) Position Rest. Potential Kinetic En. Total Energy
-4.49776e+03 2.00101e-01 -2.31992e+07 1.44043e+01 -2.31992e+07
Temperature Pressure (bar) Constr. rmsd
1.55259e-03 -1.41562e+03 5.56829e-06
My equilibration.mdp file is:
define = -DPOSRES
dt = 0.02
nsteps = 25000
nstxout = 0
nstvout = 0
nstlog = 100
nstxtcout = 100
xtc-precision = 10
rlist = 1.4
coulombtype = shift
rcoulomb = 1.2
epsilon_r = 15
vdw-type = shift
rvdw-switch = 0.9
rvdw = 1.2
tcoupl = v-rescale
tc-grps = Protein W
tau-t = 1.0 1.0
ref-t = 300 300
Pcoupl = parrinello-rahman
Pcoupltype = isotropic
tau-p = 12.0
compressibility = 3e-4
ref-p = 1.0
refcoord_scaling = all
Does anyone know how I could correct the input?
Best regards
Urszula Uciechowska
Hi Urszula Sylwia Uciechowska,
According to my knowledge the segmentation fault comes because of RAM problem.
Dear Urszula,
There are actually 5 options in my mind which you can take to avoid the problem :
1. Look at your system. Are there any atomic overlaps or unusually short distances in your system after the energy minimisation (steep,l-bfgs) ?
2. Equilibrate within many steps. Start with 10 K temperature and then increase the temperature by 50 K per step until you reach 300 K.
3. Another option : Equilibrate with a very small sigma value in the Lennard Jones potential and then increase sigma at a very low temperature 50 K for 10000 steps.
4. Try at first to equilibrate your system without constraints with constraints = none
at a very low temperature. For the production run you could add constraints.
5. Reduce the time-step at first to a very low value dt = 0.0001 so that your system cannot collapse in the beginning.
6. Be careful with the box-dimensions. With periodic boundaries you will have sometimes atom-overlaps which you cannot see at first.
Emanuel
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry
- Computational Chemistry
- Molecular Dynamics