Yes that's correct, but since that time the NEB code was implemented on
pmemd with ways to minimize the communication to the cpu for rmsd
calculations. I suggest following that code, tmd should be quite similar
and maybe even share some code sections.
Good luck!
On Wed, Apr 29, 2020, 8:00 PM Francesco Gentile <fgentile.ualberta.ca>
wrote:
> Hello everyone,
>
> I would like to perform targeted MD simulations on GPU with pmemd (Amber
> 16). I understand tMD has not yet been implemented with pmemd but I was
> thinking about modifying a bit the source code to include that feature. For
> example, by using something similar to what was done in the sander
> code. Could you let me know if this sounds reasonnable to you and if there
> is way to do that while keeping reasonnable performance on GPU? I guess tMD
> was not implemented because of RMSD calculation which may be not convenient
> for GPU architecture right?.... Do you have any advice?
>
> Thanks a lot,
>
> --
> Francesco Gentile, Ph.D.
> Vancouver Prostate Centre, Vancouver BC, Canada
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 29 2020 - 18:00:02 PDT
