Re: [ESPResSo-users] dipolar p3m on non-cubic boxes

[Valera, Manuel]

Hi
Thanks for the reply.
It would be great to get your code to see if it worth the time trying to make 
the changes. I also tried to read the file p3m-dipolar.c but as you said it is 
hard to read without some help from somebody more familiar with the code.
Thanks again and regards,
Manuel

-----Original Message-----
From: Rudolf Weeber [mailto:address@hidden 
Sent: Friday, April 19, 2013 7:58 AM
To: Valera, Manuel
Cc: address@hidden
Subject: Re: [ESPResSo-users] dipolar p3m on non-cubic boxes

Hi,
On Thu, Apr 11, 2013 at 10:33:42AM -0400, Valera, Manuel wrote:
> Hello all,
> I have a quick question, I need simulate a system with dipolar particles on 
> non-cubic boxes, Lx=Ly and Lz ~ 10*Lx. Currently dipolar p3m is limited to 
> cubic boxes and I was wondering if it is possible to extend the code to 
> non-cubic boxes. How difficult would this be and what files would we need to 
> change?
It is in principle possible. However, the code of the dipolar p3m is not easy 
to read. Therefore, it would probably take quite some time if one is not 
already familiar with the code.
I made an attempt some time ago and got a part of it done before giving up on 
it for the time.
If you're interested, I'm happy to share the git branch. 
What's done is the changes in the basic infrastructure and a rewrite of the 
influence functions based on what is written in the paper. However, something 
seems to be still wrong even with a cubic box.
Also, quite some work is still needed for the tuning.

The dipolar p3m code is in src/p3m-dipolar.c/h.

Hope, that helps.
Regards, Rudolf