AMBER Archive (2003)

Subject: Re: AMBER: more than 100,000 atoms ?

From: Robert Duke (rduke_at_email.unc.edu)
Date: Thu Aug 07 2003 - 15:33:36 CDT

David -
Sander 7 can do this, no problems. So can PMEMD, if you are mostly still in
an amber 6 context. The prmtop is not a problem but your first inpcrd file
is. Really all that is needed is the change of natom from i5 to i6; if you
are coming from tleap or xleap, I don't know how it normally handles this.
I whacked the xleap source to get it to do the right thing, myself, but that
is not a standard part of the distribution. The change I made is shown by
the following
diff (amber 6 source):
diff -tr ./leap/unitio.c /home/rduke/amber6/src.ori/leap/src/leap/unitio.c
5466c5466,5478
< FortranFormat( 1, "%5d" ); 

---
> /*
>  * Write a amber 6 inpcrd atoms line for less than 100,000 atoms, and a
>  * amber 7 inpcrd atoms line for 100,000 or more atoms.  This allows us to
>  * use PMEMD with more than 99,999 atoms.
>  */
>
>
>
>     if (iVarArrayElementCount(uUnit->vaAtoms) < 100000)
>       FortranFormat( 1, "%5d" );
>     else
>       FortranFormat( 1, "%6d" );
>

The guys I work with routinely deal with systems of over 100,000 atoms,
which is one reason I modified sander 6 into pmemd (faster, handles bigger
problems (up to about 329844 atoms)in less memory).  See the amber web page,
amber.scripps.edu,  for the link to get a copy of PMEMD.

Regards - Bob Duke

----- Original Message -----
From: "David Smith" <David.Smith_at_cup.uni-muenchen.de>
To: <amber_at_scripps.edu>
Sent: Thursday, August 07, 2003 3:44 PM
Subject: AMBER: more than 100,000 atoms ?

> Hi all,
>
> Apologies if this question has come up before but I don't remember it
> and a quick search of the archives didn't net me anything.
>
> I wanted to try a system with more than 100,000 atoms. I know it's a bit
> big but my boss is an ambitious guy.....
>
> I gave it a go and initially got very confused by getting caught by
> "Peek Ewald" in a simple sander minimization.
>
> After a bit of surprise and a few write statements I realized that NATOM
> was being read as "I5" from the CRD file. Changing it to "I6" got me
> past "Peek Ewald" but I then got stuck a bit later. The web page tells
> me that, in general, NATOM is "I6" in PARMTOP but "I5" in CRD.
>
> I am still "trapped" in Amber 6 but my guess is that these formats
> persist into Amber 7 ?
>
> Can anyone recommend a way to do such a system with AMBER (any version)?
>
> i.e. Would I have to change all the formats associated with "read (9,F)"
> statements. Is there a semi fast way to do it ? Even if I did manage
> that, would the developers imagine that something else would break and
> that I should simply abandon the whole idea.
>
> Thanks for your patience ....
>
> David.
>
>
> --
> ---------------------------------------
> Dr. David Smith
> Department of Chemistry
> Ludwig Maximilians University
> Butenandt-Str. 5-13, D-81377 Munich
> Germany
> Tel.: +49 (0)89 2180 77740
> Fax.: +49 (0)89 2180 77738
> e-mail: David.Smith_at_cup.uni-muenchen.de
> ---------------------------------------
>
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
>

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu