AMBER Archive (2007)

Subject: AMBER: Performance of polarizable model simulations

From: Jones de Andrade (johannesrs_at_gmail.com)
Date: Wed Aug 22 2007 - 23:17:52 CDT

Hi all.

I would like to make an direct question: How big is the performance loss
from a polarizable simulation to a non-polarizable one?

I would expect it to be a big hit on performance on usual implementations,
as running a whole PME to every iteration on the dipoles would be costly.
But the references on amber9 also points out that there is a "efficient" way
of doing it (still studing the paper).

The question is: how big are the gains, and the losses?

Thanks a lot to everybody in advance!

Sincerally yours,

Jones de Andrade

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