normal Segmentation faults - core dumped error

  • gypsy
  • gypsy's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
3 years 9 months ago #8602 by gypsy
Dear Martini experts,

I am trying to perform Martini CG simulations of 20 molecules of a peptide in 15x15x15 box (10mM). I am using Martini 2.1 and Gromacs 2020.1. I used the attached dynamic file, but I keep getting segmentation faults - core dumped error. Do you have any suggestions?

Thanks!!


dt = 0.02 ; 20fs timestep
nsteps = 50000000; 20fs * 50000000 = 1000ns
nstxout = 0
nstvout = 0
nstlog = 100000
nstxtcout = 100000
xtc-precision = 10
rlist = 1.4
coulombtype = PME
rcoulomb = 1.2
epsilon_r = 15
vdw-type = shift
nstlist = 100
rvdw-switch = 0.9
rvdw = 1.2
tcoupl = v-rescale
tc-grps = Protein Non-Protein
tau-t = 1.0 1.0
ref-t = 300 300
Pcoupl = parrinello-rahman
Pcoupltype = isotropic
tau-p = 40.0
compressibility = 3e-4
ref-p = 1.0
refcoord_scaling = all

Please Log in or Create an account to join the conversation.

More
3 years 9 months ago #8604 by vainikka
Replied by vainikka on topic Segmentation faults - core dumped error
The only oddity that I can spot is the high tau-p value. Do you have a specific reason for having a 40 ps coupling constant?

Also, have you performed proper equilibration steps before running a simulation with this mdp file? If I had to guess, I'd say you have high pressure oscillations, which PR barostat only worsens, leading to your system blowing up.

Please Log in or Create an account to join the conversation.

  • gypsy
  • gypsy's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
3 years 9 months ago #8610 by gypsy
Replied by gypsy on topic Segmentation faults - core dumped error
Thank you for your reply!

I tried to lower the tau-p, and to change many other options (coulombtype, vdw-type, Pcoupl, etc.) but I got still the same error.
I ran other simulations before with same parameters, and the only difference I can spot is that in those cases there were no charges (only Y, G, H, L as amino acids). Can the presence of positively charged residues influence that much the outcomes?

Thank you again.

Please Log in or Create an account to join the conversation.

More
3 years 9 months ago #8611 by vainikka
Replied by vainikka on topic Segmentation faults - core dumped error
Positive charges shouldn't cause that. How is your system behaving during the simulation? Does the segmentation fault occur immediately, or after some time?

Usually these errors are caused by the system 'blowing up' ( www.gromacs.org/Documentation_of_outdate...rminology/Blowing_Up ). Have you equilibrated the system properly? Have you tried a smaller time step?

Please Log in or Create an account to join the conversation.

  • gypsy
  • gypsy's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
3 years 9 months ago #8612 by gypsy
Replied by gypsy on topic Segmentation faults - core dumped error
The segmentation fault occurs pretty soon (within 10min).

I have used this parameters for the equilibration (slash stays for other conditions I tried):


;define = -DPOSRES
dt = 0.02
nsteps = 50000
nstxout = 0
nstvout = 0
nstlog = 1000
nstxtcout = 1000
xtc-precision = 10
rlist = 1.4
coulombtype = PME/ reaction-field (vdw-modifier = Potential-shift-verlet; verlet-buffer-tolerance= 0.005)
rcoulomb = 1.2
epsilon_r = 15
vdw-type = shift/ cutoff
rvdw-switch = 0.9
rvdw = 1.2
tcoupl = v-rescale
tc-grps = Protein Non-Protein
tau-t = 1.0 1.0
ref-t = 300 300
Pcoupl = parrinello-rahman/Berendsen
Pcoupltype = isotropic
tau-p = 12.0
compressibility = 3e-4
ref-p = 1.0
refcoord_scaling = all


No I haven't tried to decrease dt. I will!

Please Log in or Create an account to join the conversation.

Time to create page: 0.107 seconds