AMBER Archive (2006)

Subject: Re: AMBER: ptraj vector

From: Thomas Cheatham (tec3_at_utah.edu)
Date: Tue Jun 20 2006 - 16:59:50 CDT

> > ==========================
> > vector v1 @12 [corr] @13
> > analyze timecorr vec1 v1 tcorr 2000 dplr out cn.cf
> > go
>
> Weird.. I use the "vector" command all the time, but I always use the
> ":2_at_N" type syntax, (i.e. residue number, atom name). It might be worth
> playing with alternatives in the atom-mask syntax -- at least you might get a
> different error message!
>
> ...just grasping at straws here....dac

The problem is that with the "corr" option to vector, per frame data is
not allocated or saved, so attempts to write it out or access it fail. I
can run recent versions of ptraj w/out core dump, but no result is
supplied as no vector data as a function of frame was ever saved. This is
kind of a NULL result (i.e. a lot of stuff is calculated and then ignored
as far as I can tell).

For me to get a better handle, I have to dissect exactly what that "corr"
option is doing with "vector" and I will try to do this over the next few
days (in consultation with Holger Gohlke when I see him this weekend :-).

Standard correlations of vectors (without the "corr" keyword in the vector
command) seem to act as expected...

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