Date: Thu, 8 Mar 2001 10:53:51 -0800
From: David Case
Subject: Re: sander VS sander_classic

On Thu, Mar 08, 2001, Khatcharin Siriwong wrote:
>
> I obtained the SANDER result which quite different from SANDER_CLASSIC
> result and
> running SANDER toke cpu time more than running SANDER_CLASSIC.
>

Please see section 6.3 of the Users' Manual for a fairly extensive
discussion of the differences between the PME methods in sander and
sander_classic, including instructions about how to get equivalent
results in the two programs.

Please note, however, that we strongly recommend using sander for PME
calcluations (so much so that we will probably disable PME in future
versions of sander_classic.) The extra time cpu time is generally
"worth it" in terms of quality and stability of the simulations (my
opinion, but not mine alone.)

Once again, here is my recommendation:

(1) Use sander for PME and generalized Born simulations

(2) Use sander_classic for everything else.

(3) Please avoid violating rules (1) and (2) until/unless you are a
fairly experienced Amber user willing to carefully investigate what is
going on.....

....hope this helps...dac