Re: NAMD parameters +mergegrids +ignoresharing

From: Aron Broom (broomsday_at_gmail.com)
Date: Fri Nov 30 2012 - 09:41:04 CST

Hi Norman,

I don't have detailed information about +mergegrids. It was recommended as
an option to speed up my GBIS GPU calculations by one of the NAMD team, and
it worked quite well (~15% increase in performance for a 2k atom GBIS
system on a single GPU). It is not an allowed parameter for CPU
calculations. My initial thought was that it merged the patches in
simulation coordinate space, but I guess it says grids, so I've no real
understanding of what is happening. Apparently it won't hurt accuracy, but
for larger systems it will start to reduce performance.

~Aron

On Fri, Nov 30, 2012 at 1:49 AM, Norman Geist <
norman.geist_at_uni-greifswald.de> wrote:

> Hi people,****
>
> ** **
>
> could someone please explain this two parameters for me, I’ve not found
> any information about it. ****
>
> Even no idea where you people got them from, just saw you using them.****
>
> ** **
>
> Regards****
>
> ** **
>
> ~Norman****
>

-- 
Aron Broom M.Sc
PhD Student
Department of Chemistry
University of Waterloo

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:22:19 CST