Hi,
The way that I would approach this is to use cpptraj to generate
matrices for certain parts of the trajectory, then run 'diagmatrix'
with 'thermo' and 'outthermo' on each. For example:
matrix name M_100ns start 1 stop 100000 mwcovar ...
matrix name M_200ns start 1 stop 200000 mwcovar ...
...
diagmatrix M_100ns thermo outthermo M_100ns.thermo.dat temp 300
diagmatrix M_200ns thermo outthermo M_200ns.thermo.dat
...
The MPI version of cpptraj can help you generate the matrices faster
by dividing the trajectory read among threads but you should do some
benchmarking on a subset of frames first to see when your IO becomes
saturated. Note that the 'diagmatrix' command is not MPI parallelized
and will be executed by the main thread only.
Hope this helps,
-Dan
On Tue, May 24, 2016 at 12:16 PM, SHAILESH KUMAR <shaile27_sit.jnu.ac.in> wrote:
> Hello,
>
> I am trying to estimate entropic contribution to the binding free energy
> for binding of a small drug-like molecule to 286 residue receptor using QHA
> with MMPBSA.py. I am using AmberTools15. My simulation run is 1
> micro-second, frames saved every 1ps, and I am able to get entropy
> estimates from entire 1 micro-sec, but I am interested in convergence of
> these estimates as well, are there any ways to output QHA values say after
> (100, 200, 300, 400 .. 900ns) so that convergence can be checked.
>
> My problem is that running this analysis takes me ~20 days on 48-core
> SMP-machine for 1 micro-second simulation. this way being able to run in
> above mentioned way seems unfeasible.
>
> Any suggestion in this regard would be very helpful.
>
> Thanks
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
-------------------------
Daniel R. Roe, PhD
Department of Medicinal Chemistry
University of Utah
30 South 2000 East, Room 307
Salt Lake City, UT 84112-5820
http://home.chpc.utah.edu/~cheatham/
(801) 587-9652
(801) 585-6208 (Fax)
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue May 24 2016 - 12:30:03 PDT
