I am running the two CPU-based MD jobs using the same node of a linux based computer cluster (CentOS Linux release 7.7.1908) (CPU: Intel(R) Xeon(R) Gold 6152).
Run Command:
mpirun -np 4 sander.MPI -O -i mdsim.in -o mdsim.out -p an_mol.prmtop -c saved.rst -x mdsim.crd -e mdsim.en -r mdsim.rst
Abdullah
________________________________
From: Carlos Simmerling <carlos.simmerling.gmail.com>
Sent: Thursday, January 23, 2020 12:43 AM
To: AMBER Mailing List <amber.ambermd.org>
Subject: Re: [AMBER] Wanting to get synchronized trajectory files
this would only be possible on identical hardware - can you tell us more
about how you are running the 2 jobs?
On Wed, Jan 22, 2020 at 10:05 AM Abdullah Bin Faheem <
abdullahbinfaheem.hotmail.com> wrote:
> Thank you for your response. While using 100 steps produces the same
> coordinates, noticeable deviations in the coordinate file begin to appear
> after 3000 steps. Is it not possible to run longer trajectories? Say 50 or
> 100 ps without experiencing changes.
>
> Abdullah
> ________________________________
> From: David Case <david.case.rutgers.edu>
> Sent: Wednesday, January 22, 2020 10:09 PM
> To: AMBER Mailing List <amber.ambermd.org>
> Subject: Re: [AMBER] Wanting to get synchronized trajectory files
>
> On Wed, Jan 22, 2020, Abdullah Bin Faheem wrote:
> >
> >Is it possible to get the same output trajectory file if I use the same
> >topology and restart file?
> >Meaning if I run two MD runs in parallel, can I get the same xyz (or
> >atleast very close) coordinates being written in the output coordinate
> >file if I use the same restart and topology file? This is the input file
> >I used initially (I am currently using AMBER 12):
>
> What happenes when you tryrunning (say) two 100-step simulations?
> Don't be afraid to experiment. Be sure to set ig (I'm not sure what the
> default value for that was way back in 2010.)
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jan 22 2020 - 08:30:02 PST
