Den 2018-08-13 kl. 11:49, skrev Chi Yuen Pak:
Dear Prof. van der Spoel,
Thanks for your reply. I tried touse the following commands (2 OpenMP threads):
gmx_mpi grompp -f eq.mdp -c em.gro-p swm4-dp.top -r posre.gro -o eq.tpr
mpirun -np 1 gmx_mpi mdrun -ntomp 2-deffnm eq
Did you try with thread-mpi?
mpirun -np 1 gmx_mpi mdrun -ntomp 1 -ntmpi 2
but this gave me the following errormessage in GROMACS 2018.2:
-------------------------------------------------------
Fatal error:
OpenMP threads have been requestedwith cut-off scheme Group, but these are only
supported with cut-off schemeVerlet
--------------------------------------------------------
I can successfully runthe simulation only when I set “–ntomp 1” with “Group” in nonPBC (since “Verlet” as the cutoff scheme is only supported in full PBC or pbc=xy)
The mdp, topology, and logfiles are linked at the end for your reference.
In addition, I have alsotried –pd option with GROMACS 4.6.7, but as far as I know, particledecomposition has not been supported since GROMACS 5.0.
There is another problemwith version 4.6. It seems that the flat-bottom restraint is also not supportedbefore version 5. It was reported that an error message would appear ifflat-bottom restraint was used in the version before GROMACS 5.0 (Ref.: https://redmine.gromacs.org/issues/1794). Therefore, may I ask how you carried out thesimulations with flat-bottom restraints in GROAMCS 4.6 in your 2011 paper?
I guess we implemented the flat-bottom restraints in 4.6.7 but it was
only included in grmoacs from 5.0.
Thanks again for yourhelp.
Best,
Chi Yuen
mdp:https://www.dropbox.com/s/xq6s63x4j30hl1h/eq.mdp.pdf?dl=0
topology:https://www.dropbox.com/s/0gj8ot2m6k8b019/swm4-dp.top.pdf?dl=0
logfile when setting –ntomp to be 1: https://www.dropbox.com/s/p2pi2yvoo9x4d39/eq-1ntomp.log.pdf?dl=0
log file when setting –ntomp to be 2: https://www.dropbox.com/s/zn2pqr6qnrf192f/eq-2ntomp.log.pdf?dl=0
| | Virus-free. www.avast.com |
David van der Spoel <sp...@xray.bmc.uu.se> 於 2018年08月12日 (週日) 2:36 PM 寫道﹕
Den 2018-08-11 kl. 16:33, skrev Chi Yuen Pak:
Dear GROMACS users,
I would like to ask a question concerning about the simulation of a water
droplet, with SWM4-DP polarizable water model in nonperiodic boundary
conditions. The GROMACS version I use is 2018.2. I have been only able to
successfully run the simulation with only “Group” as the cutoff scheme in non
PBC. Multiple OpenMP threads are not supported with “Group”. Also, shell
particles are not implemented with domain decomposition, so I have not been
able to run the simulation with more than a single core. The following commands
are what I use (with thread-MPI)gmx_mpi grompp -f eq.mdp -c em.gro -p
swm4-dp.top -r posre.gro -o eq.tprmpirun -np 1 gmx_mpi mdrun -ntomp 1 -deffnm eq
Are you sure it is not using more than one thread? What does it say in
the log file?
In the paper “Atomistic simulation of ion solvation in water explains surface
preference of halides” by Prof. David Van der Spoel in 2011 (doi:
10.1073/pnas.1017903108), they also used the SWM4-DP water model in non PBC (no
cutoffs) in GROMACS. I can only guess they used multiple cores for their
simulations, or did they? Does anyone know if it is possible to run multiple
cores to simulate SWM4-DP in non PBC?
That was run with gromacs 4.6 using the -pd option.
You could do that as well of course.
Attached are the mdp setting and topology for your reference.Thanks.
Best,Chi Yuen
mdp setting: https://www.dropbox.com/s/e1sord21slgn084/eq.mdp.pdf?dl=0
topology: https://www.dropbox.com/s/0gj8ot2m6k8b019/swm4-dp.top.pdf?dl=0
| | Virus-free. www.avast.com |
--
David van der Spoel, Ph.D., Professor of Biology
Head of Department, Cell & Molecular Biology, Uppsala University.
Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
http://www.icm.uu.se
--
Gromacs Users mailing list
* Please search the archive at
http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting!
* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a
mail to gmx-users-requ...@gromacs.org.
