AMBER Archive (2005)

Subject: Re: AMBER: amber8 parallel sander

From: Robert Duke (rduke_at_email.unc.edu)
Date: Thu Sep 22 2005 - 16:47:10 CDT


Your information is old. Even sander is doing better than this, and pmemd
holds its own out there for a range of processors most folks have access to.
Look at the benchmarks on amber.scripps.edu.
Regards - Bob

----- Original Message -----
From: "Peter Gannett" <pgannett_at_hsc.wvu.edu>
To: <amber_at_scripps.edu>
Sent: Thursday, September 22, 2005 5:15 PM
Subject: Re: AMBER: amber8 parallel sander

> Kathleen:
>
> Yes. But, Amber users should not find this to really be a problem as
> NAMD can use amber's force field and I've not seen any reports of
> significant differences between NAMD with amber's ff and Amber. Also,
> from what I've seen reported (amber reflector, other various websites,
> though some possibly biased like if they are from the makers of NAMD or
> Amber), the switching point is somewhere between 4-8, 8 seems to come up
> consistently, and by 16 processors, you don't want to be using amber.
>
> Pete
>
>>>> kathleen_at_massivelyparallel.com 09/22 12:44 PM >>>
> Peter: So are you saying that scaling efficiencies beyond 8 CPUs are
> better
> achieved with NAMD? In other words, is 8 CPUs the limit for most
> effiencient scaling on AMBER? I'd like to know at which point AMBER
> users
> find they need to switch to NAMD or some other MD package. -Kathleen
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>

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