AMBER Archive (2009)

Subject: RE: [AMBER] Is the result of sander same as that of pmemd?

From: Ross Walker (
Date: Sat Jun 13 2009 - 11:12:51 CDT

Hi Catein,

> What is the topical values of ntrp ntwx and ntwr for long MD, as the
> mdcrd files and rstrt will be getting very large the cannot be handle
> easily? Could you mind to give me some suggestions?
> I noted ntrp=100, ntwx=100 and ntwr=1000 was used in the tutorials.
> Can we increased them further for 3 ns simulations?

You can set these to any values you want, of course under the caveat that
you are effectively throwing data away so it really comes down to what you
want to learn and the timescales that occurs on. An understanding of
statistical mechanics will help you here.

Typically I use values in the range of ntwx = 1000 and ntpr = 1000 with
ntwr=10000. You might also want to consider using netcdf binary trajectories
(check the manual) since this will reduce the file sizes somewhat. Beyond
that, buy lots of disk. Note you can always post process the files to strip
out additional frames if you want. What you can't do of course is put frames
back in since this information was thrown away.

All the best

|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- |
| | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

AMBER mailing list