From: root (root_at_liposome.genebee.msu.su)
Date: Thu Jan 08 1998 - 12:44:37 CST

This should be the last bit about trajectory conversion.

---------- Forwarded message ----------
Date: Thu, 8 Jan 1998 12:49:08 -0500 (EST)
From: Rick Venable <rvenable_at_deimos.cber.nih.gov>
To: Leif Laaksonen <laaksone_at_csc.fi>
Cc: Konrad Hinsen <hinsen_at_ibs.ibs.fr>, root_at_liposome.genebee.msu.su,
    chemistry_at_infomeister.osc.edu
Subject: Re: CCL:md rookie update

On Thu, 8 Jan 1998, Leif Laaksonen wrote:
> If you aren't worried about loosing the last decimal of the accuracy you
> can use my small FORTRAN program for CHARMM trajectories for the
> formatting/unformatting at:
> http://laaksonen.csc.fi/gopenmol/distribute/linux/charmmtrj.f
> I mostly use the program to move the trajectories between different
> types of machines.

Of course, if you're actually using CHARMM, you can use the built-in FORMAT
and UNFORMAT commands to move CHARMM trajectories across machines types
(whether you call them .dcd or .trj files). It's a feature that's been
available for well over a decade, since the differences in machine dependent
floating point representations were even greater in the past. It allows
the user some control over the precision, by changing some of the scaling and
the Z format specs (Fortran zoned-hexadeciamal format).

I should add that, besides big-endian and little-endian IEEE f.p.
representations, there can be file structure differences between platforms as
well. The Fortran standard does not define the structure of a binary file on
disk, only the results of READ and WRITE operations; the details can be
machine dependent.

--
Rick Venable                  =====\     |=|    "Eschew Obfuscation"
FDA/CBER Biophysics Lab       |____/     |=|
Bethesda, MD  U.S.A.          |   \    / |=|  ( Not an official statement or
rvenable_at_deimos.cber.nih.gov  |    \  /  |=|    position of the FDA; for that,
http://nmr1.cber.nih.gov/           \/   |=|    see   http://www.fda.gov  )
http://www.erols.com/rvenable