Metadynamics run using GROMACS/2020.4-intel-2020u2-PLUMED-2.7.0 fails

User discussions
1

GROMACS version: 2020.4
GROMACS modification: Yes/No (plumed enabled)
Here post your question:
I am running a Metadynamics simulation with multiple replicas using gromacs (plumed enabled version). It was running fine for two days (22ns) but then I wanted to continue the run beyond 2 days using the checkpoint file. Yet every time it gives me an error along these lines:

Program: gmx mdrun, version 2020.4-MODIFIED
Source file: src/gromacs/ewald/pme_redistribute.cpp (line 305)
MPI rank: 34 (out of 400)

Fatal error:
1711 particles communicated to PME rank 2 are more than 2/3 times the cut-off
out of the domain decomposition cell of their charge group in dimension x.
This usually means that your system is not well equilibrated.

For more information and tips for troubleshooting, please check the GROMACS
website at Errors - Gromacs

mdp file:
#################

integrator = md
dt = 0.002
nsteps = -1
nstxout-compressed = 5000
compressed-x-precision = 1000
nstxout = 50000
nstvout = 50000
nstfout = 0
nstcalcenergy = -1
nstenergy = 5000
nstlog = 10000
;
cutoff-scheme = Verlet
nstlist = 20
rlist = 1.2
coulombtype = pme
rcoulomb = 1.2
vdwtype = Cut-off
vdw-modifier = Force-switch
rvdw_switch = 1.0
rvdw = 1.2
;
pme_order = 4
ewald_rtol = 1e-05
ewald-rtol-lj = 0.001
lj-pme-comb-rule = Geometric
ewald_geometry = 3d
epsilon_surface = 0
fourierspacing = 0.12
;
tcoupl = V-rescale
nsttcouple = -1
tc_grps = Protein non-Protein
tau_t = 1.0 1.0
ref_t = 310 310
;
pcoupl = Parrinello-Rahman
pcoupltype = isotropic
tau_p = 1.0
compressibility = 4.5e-5
ref_p = 1.0
;
constraints = h-bonds
constraint_algorithm = LINCS
continuation = no
lincs-order = 6
lincs-iter = 2
lincs-warnangle = 30
;
nstcomm = 25
comm_mode = linear
comm_grps = SYSTEM
;
pbc = xyz
;
refcoord_scaling = com
gen-seed = -1
ns_type = grid

#############

Job script:
###############
#SBATCH -J mnm_nmr # Name of the job
#SBATCH -n 400 #–gres=gpu:1
#SBATCH -N 9 # Number of nodes
#SBATCH -t 2-00:00:00 # Runtime in HH:MM:SS
#SBATCH --mem-per-cpu=2GB # Memory requested in MB (see also --mem-per-cpu)
#SBATCH -o mnm_nmr_cont%j.out # File to write STDOUT, %j=jobid
#SBATCH -e mnm_nmr_cont%j.err # File to wrote STDERR, %j=jobid
#SBATCH --mail-type=ALL # Send email when job starts, ends, fails, etc
\

Load the “gromacs” module

module load GROMACS/2020.4-intel-2020u2-PLUMED-2.7.0

Metadynamics Metainference

mpirun -np 400 gmx_mpi mdrun -v -plumed plumed_mnm_v5.dat -deffnm mnm -multidir {0…19} -cpi mnm.cpt -maxh 47 -cpt 40 -tunepme yes
############

I have tried troubleshooting using different things but it’s just not working so I decided to reach out for some expert help. I would really appreciate guidance on how to handle this error.

P.S. the system looks fine.

regards,
Zahra

2

Hi,
I believe you have already performed some equilibration steps prior to the current Parrinello-Rahman simulation, however, according the ‘‘continuation = no’’ in your mdp file you are not basically using the equilibrated system. The ‘‘continuation = no’’ means generation of the initial velocity of the particles from scratch.
So, try ‘‘continuation = yes’’ and ‘‘gen-vel = no’’ and use (gmx gromp … -t *.cpt) the checkpoint file of the previous simulation.

Regards,
Salman

3

Thank you for your recommendation Salman!
I will try to follow it and see how it goes.

regards,
Zahra

6

Hello,

I tried the advised method but it did not help me.

I used the following commands:

gmx grompp -f metad-metai4_v2.mdp -p topol.top -t mnm.cpt -c mnm.tpr -o mnm_v2.tpr

And then used this subsequent mdrun command:

mpirun -np 400 gmx_mpi mdrun -v -plumed plumed_mnm_v5.dat -deffnm mnm -s mnm_v2.tpr -multidir {0…19} -cpi mnm.cpt -maxh 47 -cpt 40 -tunepme yes

Maybe I am still doing something wrong because the simulation stopped after 3 minutes into the run. Is there anything else I can check.

regards,
Zahra

7

HI,
I tried the advised method but it did not help me.

I used the following commands:

gmx grompp -f metad-metai4_v2.mdp -p topol.top -t mnm.cpt -c mnm.tpr -o mnm_v2.tpr

And then used this subsequent mdrun command:

mpirun -np 400 gmx_mpi mdrun -v -plumed plumed_mnm_v5.dat -deffnm mnm -s mnm_v2.tpr -multidir {0…19} -cpi mnm.cpt -maxh 47 -cpt 40 -tunepme yes

Maybe I am still doing something wrong because the simulation stopped after 3 minutes into the run. Is there anything else I can check.

8

reply
reply1414×766 148 KB

9

Hi Zahra,
If it is not an issue with your specific system and the metadynamics simulation which I am not that much into it, then it all gets back to have the system well energy-minimized and well-equilibrated.

Some potential solutions could be:

  1. Please first try to run the system for one or two ns using a less fancier pressure coupling algorithm likes the Berendsen and then switch to Parrinello.

  2. Sometimes changing the maximum considered distance for the bonding interaction would give some more room, so, no hurt of you test something like -rdd 1.2 -dds 0.9999 or around in your gmx mdrun.

Cheers,
Salman

10

Thank you for your help!
Unfortunately, this problem still persists after equilibration so I am thinking that it might be because of a bug of some sorts.

regards,
Zahra