[gmx-users] Re: Re:Which kind of box?

[Tsjerk Wassenaar]

Xie Yinghong,

It's okay, apparently I'm failing in making myself clear. I've dug up this
image which visualizes the interconversion of box types. You see a molecule
(blue) in a hexagonal unit cell (yellow). The centres of these unitcells are
connected by the black lines, which form the lattice. One cell, contained by
four lattice points, is exactly equal in volume and composition as the
original hexagonal unit cell. I hope this is clear now.

But we've wandered a long way, and if I'm not mistaken your simulations
crashed, and you thought it had to do with the box. Now, understanding how
periodic boundaries work, the infinite simulation system and the
interconversion of box types is important in MD, I think it does not solve
your problem. It would only clear you of the misconception that it had to do
with your box being triclinic. The cause of the crash is not related to the
nature of the box, unless you've set it to be too small to contain your
entire system.

Tsjerk

On 10/15/05, Yinghong <xieyh at hkusua.hku.hk> wrote:
>
> Dear Dr. Tsjerk:
>
>  Firstly, sorry to occupy you so much time to deal with my questions,
> thanks again.
>
>  But, I do not know why you always mentioned about a hexagonal prism and
> why I need to make such a transformation from parallelogram to hexagon.
>
>  The box I need is only a parallelogram one. As a fact, I can get such a
> box through defining "angles=90 90 120" in editconf command, and a
> parallelogram shape box can be seen by trjconv. So, I can not understand why
> hexagon is always mentioned?
>
>  Thanks and regards,
>
>  Yinghong
>
> Hong Kong University
>
>  >You just don't get it... Now find and read the paper of Bekker and do
> the
> >excercise from the previous mail. The answers are:
>
> >1. Your parallelogram can be transformed to a single hexagonal prism with
> >exactly the same lattice vectors, volume and content. One to one.
>
> >2. I think that when setting the box with -vectors and -angles, -bt tric
> is
> >implicit, but it doesn't hurt anyway.
>
> >Tsjerk
>
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
>
>


--

Tsjerk A. Wassenaar, M.Sc.
Groningen Biomolecular Sciences and Biotechnology Institute (GBB)
Dept. of Biophysical Chemistry
University of Groningen
Nijenborgh 4
9747AG Groningen, The Netherlands
+31 50 363 4336
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20051014/0695fcb4/attachment.html>