Hi,
On Wed, Jun 2, 2021 at 11:27 AM Jenny 148 <jenny.rs140.gmail.com> wrote:
> In cpptraj would it be enough to give the line
> cpptraj -i traj.in
This should work fine. If you have access to multiple nodes you may
get a speedup by using cpptraj.MPI. I've typically found good results
on HPC systems by using cpptraj.MPI.OMP (MPI + OpenMP enabled) and
using as many processes as there are CPU sockets and using openmp
threads for the remaining cores. So for example, for 2 nodes where
each node has 2 sockets and each socket has 12 cores, I would use 4
MPI processes, 12 threads per process.
Hope this helps,
-Dan
>
> With the traj.in containing the following lines:
> parm prot.prmtop
> trajin prod.nc
> reference equil.rst7
>
> rms reference out ligand_500ns.rms :1
> --
> Jenny R.S
> _______________________________________________
> 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 Jun 02 2021 - 11:00:02 PDT
