AMBER Archive (2005)

Subject: Re: AMBER: Problem with computing Resp charges using RED

From: FyD (fyd_at_u-picardie.fr)
Date: Sat Jan 01 2005 - 13:51:12 CST


Dear Cenk Andac,

> I have found out that RED calculations take
> considerable amount of hard-disk space while GAMESS is
> executed. My molecule is about 300 Daltons.
> I noticed that RED stopped running because my
> hard-disk ran out of disk-space. RED has consumed all
> my available disk space (about 2.2 GBytes!). Is this
> the way RED usually works?

R.E.D. is just an interface to GAMESS/Gaussian and RESP. So what you described
has nothing to do with R.E.D. but is just a feature of GAMESS/Gaussian...

> Do I need to purchase an
> extra hard-disk for RED or are there other ways to
> reduce hard-disk space consumption while execution of
> RED?

Yes you can but usually if your molecule is too big you can/should split it into
several pieces. This means that instead of calculating RESP charges for one
unique unit you do it for 'X' units. For instance, a peptide bond can be used
to split your molecule into tow parts... Using once again intra-molecular
restraints in RESP...

An example is below:
Instead of calculating the RESP charges for the whole molecule "R1-NHCO-R2" you
can do it for:
R1-NH-'COMe' & 'MeNH'-CO-R2 using intra-molecular restraints (= 0) for '-COMe'
and 'MeNH-'...
This _will_ be automatically implemented in R.E.D. III.

> After excavating the GAMESS directory, I noticed that
> the culprit is a bulky file (about 2.2 GByte)
> generated in a temp directory of GAMESS while RED is
> running.

Check also if your cpu is a 64bits or a 32bits to overcome the 2GB file size
limit (for 32bit cpu) for your QM temporary files...

> I will appreciate your suggestions ragarding the
> extreme hard-disk consumptions during execution of
> RED.

Splitting your molecule in several pieces is an answer and it will save you a
loooot of computer time in particular for geometry optimization step...

Regards, Francois

  F.-Y. Dupradeau
DMAG EA 3901 & Faculte de Pharmacie, Amiens, France
The Scripps Research Institute, La Jolla, CA, USA
       --
http://www.u-picardie.fr/labo/lbpd/FyD.htm
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu