[gmx-users] Octahedral minimization problem

[Ahmed Mashaly]

Hi

Two solvated systems with amber were converted to .gro with acpype. one is box the other is octahedral


The box can be minimized in cluster with this em.mdp file


integrator = steep 
emtol = 100.0 
emstep = 0.01 
nsteps = 50000 

nstlist = 1 
cutoff-scheme 
ns_type = grid 
coulombtype = PME 
rcoulomb = 1.0
rvdw = 1.0 
pbc = xyz 


The box can be minimized, but at the end I had 

Steepest Descents converged to Fmax < 100 in 2706 steps
Potential Energy = -2.7121270e+06
Maximum force = 9.3153786e+01 on atom 163971
Norm of force = 1.7101741e+00
Simulation ended prematurely, no performance report will be written.

as I understood from other messages in archive this premature is not a problem as it ends with good pot energy and max force

While in case of octahedral, I got this:

Steepest Descents:
Tolerance (Fmax) = 1.00000e+02
Number of steps = 50000
Step= 0, Dmax= 1.0e-03 nm, Epot= 8.27123e+15 Fmax= 1.65730e+17, atom= 103876
Step= 1, Dmax= 1.0e-03 nm, Epot= 4.22395e+15 Fmax= 6.54268e+16, atom= 103876
Step= 2, Dmax= 1.2e-03 nm, Epot= 1.96761e+15 Fmax= 2.32216e+16, atom= 103876
Step= 3, Dmax= 1.4e-03 nm, Epot= 8.64595e+14 Fmax= 7.37977e+15, atom= 103876
Step= 4, Dmax= 1.7e-03 nm, Epot= 3.77835e+14 Fmax= 2.09380e+15, atom= 103876
Step= 5, Dmax= 2.1e-03 nm, Epot= 1.49451e+14 Fmax= 5.29119e+14, atom= 103876
Step= 6, Dmax= 2.5e-03 nm, Epot= 5.37730e+13 Fmax= 1.41647e+14, atom= 34363
Step= 7, Dmax= 3.0e-03 nm, Epot= 1.82718e+13 Fmax= 3.59846e+13, atom= 199780
Step= 8, Dmax= 3.6e-03 nm, Epot= 6.05245e+12 Fmax= 7.92220e+12, atom= 134203
Step= 9, Dmax= 4.3e-03 nm, Epot= 1.72087e+12 Fmax= 1.62912e+12, atom= 72328
Step= 10, Dmax= 5.2e-03 nm, Epot= 4.91989e+11 Fmax= 3.31401e+11, atom= 106294
Step= 11, Dmax= 6.2e-03 nm, Epot= 1.59698e+11 Fmax= 7.11892e+10, atom= 136966
Step= 12, Dmax= 7.4e-03 nm, Epot= 5.16370e+10 Fmax= 1.55656e+10, atom= 14985
Step= 13, Dmax= 8.9e-03 nm, Epot= 1.66064e+10 Fmax= 6.35415e+09, atom= 14986
Step= 14, Dmax= 1.1e-02 nm, Epot= 8.82293e+09 Fmax= 3.16163e+09, atom= 14985
Step= 15, Dmax= 1.3e-02 nm, Epot= 4.84420e+09 Fmax= 8.10766e+08, atom= 14986
Step= 16, Dmax= 1.5e-02 nm, Epot= 1.68188e+09 Fmax= 2.50214e+08, atom= 14985
Step= 17, Dmax= 1.8e-02 nm, Epot= 7.22330e+08 Fmax= 3.72832e+07, atom= 14986
Step= 18, Dmax= 2.2e-02 nm, Epot= 1.55106e+08 Fmax= 8.66899e+06, atom= 17209
Step= 19, Dmax= 2.7e-02 nm, Epot= 7.04906e+07 Fmax= 4.10104e+06, atom= 152596
Step= 20, Dmax= 3.2e-02 nm, Epot= 4.15026e+07 Fmax= 1.23764e+06, atom= 152596
Step= 21, Dmax= 3.8e-02 nm, Epot= 1.95151e+07 Fmax= 8.10870e+06, atom= 164668
Step= 22, Dmax= 4.6e-02 nm, Epot= 1.78878e+07 Fmax= 9.02967e+05, atom= 164668
Step= 23, Dmax= 5.5e-02 nm, Epot= 1.14131e+07 Fmax= 5.10415e+06, atom= 164668

step 24: One or more water molecules can not be settled.
Check for bad contacts and/or reduce the timestep if appropriate.

step 24: One or more water molecules can not be settled.
Check for bad contacts and/or reduce the timestep if appropriate.
Wrote pdb files with previous and current coordinates
Wrote pdb files with previous and current coordinates
Fatal error in MPI_Sendrecv: Message truncated, error stack:
MPI_Sendrecv(259).................: MPI_Sendrecv(sbuf=0x7ffed5bfe230, scount=8, MPI_BYTE, dest=8, stag=0, rbuf=0x7ffed5bfe238, rcount=8, MPI_BYTE, src=6, rtag=0, MPI_COMM_WORLD, status=0x7ffed5bfdf90) failed
MPIDI_CH3U_Receive_data_found(131): Message from rank 6 and tag 0 truncated; 17016 bytes received but buffer size is 8


>From the checked archive I know people had similar problems will be related to topology and atomic clashes, but I don`t have any and some meaningless pdb files were produced with the name of this step. but when I tried on my laptop, the same water error appeared, but the minimization process continued with pdb (s) created and at the end I got this with gro file and everything and this log:


Energy minimization has stopped, but the forces have not converged to the
requested precision Fmax < 100 (which may not be possible for your system).
It stopped because the algorithm tried to make a new step whose size was too
small, or there was no change in the energy since last step. Either way, we
regard the minimization as converged to within the available machine
precision, given your starting configuration and EM parameters.

Double precision normally gives you higher accuracy, but this is often not
needed for preparing to run molecular dynamics.
You might need to increase your constraint accuracy, or turn
off constraints altogether (set constraints = none in mdp file)

Steepest Descents converged to machine precision in 13158 steps,
but did not reach the requested Fmax < 100.
Potential Energy = -3.8879410e+06
Maximum force = 1.8154966e+03 on atom 21160
Norm of force = 5.8444543e+00

Simulation ended prematurely, no performance report will be written.



Later when I tried to run it in the cluster, but with only one cpu instead of many (48, 10, 5 were tried and got the same error), but with only one cpu it worked as the same on my laptop. so what is the problem with MPI in the case of octahedral box? and can I consider this minimization with one processor is enough 


Kind Regards,
Ahmed