Hi All,
With regards to the tutorials in the beginning of the DNA tutorial 1 I use a
frequency of 100 x 2fs for the first two simulations. Mainly because these
are small gas phase simulations that are designed to blow up and I want the
user to see this.
In the final stage of the tutorial where the users run a simulation that
converts A-DNA into B-DNA then I use 250x2fs which is perhaps a little too
often although this is only a 1.8ns run. I could perhaps add a discussion
here about suitable values for ntwx.
The other tutorial that does long MD runs is tutorial 8 which is TRPcage
folding here during the heating phase I use 50x0.5fs for mdcrd but this is
because the initial structure is very strained and I find if it is going to
go bang it will do so during the initial heating (40ps or so) and having a
frequent mdcrd dump can show up potential problems. When I move to the
production which are 5ns blocks I use 500x2fs but this was mainly because
the system only has 304 atoms total and being GB the time per step is quite
high. Again perhaps I'll add a note here saying that if this was an explicit
water simulation then a value of 1000 or so for ntwx would be more
appropriate.
On another note, I think it might be a good idea to introduce users to using
PMEMD for calculations it supports although the question is where to put
this such that it doesn't overly confuse people. I am currently writing a
tutorial that shows how to install amber on linux and windows so perhaps I
will add it here with an explanation.
All the best
Ross
/\
\/
|\oss Walker
| HPC Consultant and Staff Scientist |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
|
http://www.rosswalker.co.uk | PGP Key available on request |
Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.
> -----Original Message-----
> From: owner-amber-developers.scripps.edu
> [mailto:owner-amber-developers.scripps.edu] On Behalf Of Robert Duke
> Sent: Friday, May 05, 2006 07:34
> To: amber-developers.scripps.edu
> Subject: Re: amber-developers: Troubles at PSC
>
> Adrian -
> Well, actually there are lots of users out there who probably
> started using
> amber before Ross' excellent tutorials were available. I
> have looked at
> some, but not all of the tutorials, and I have nothing but
> praise. I don't
> recollect what they say about ntwx settings, and have not
> plowed through
> them all to see. One problem with tutorials though, is
> typically they are
> trying to show how to use features with a toy problem so the user can
> produce results in a reasonable time. SO it would be
> understandable if a
> tutorial generated a short trajectory using higher frequency
> snapshots.
> Maybe the tutorials do cover this, but as I say, I am sure
> not everyone has
> been through every page. Okay, back to the manual. I
> perused it quickly.
> It has at least two early examples with ntwx set to 100. It
> has one example
> with it set to some huge number, rather than just not setting
> it or setting
> it to 0 (so this can produce confusion). It has a couple of
> examples of it
> set to 500 for PB (that is too high a frequency unless dt =
> .002 - I don't
> remember what it was). It has one charming example of a
> MINIMIZATION being
> run with ntwx and a few other things that make no sense being set to
> moderate numbers. I am a details freak, and I spot
> inconsistencies pretty
> readily. Other people copy inconsistencies rather readily.
> If you go out
> on the amber mail reflector archive and search for ntwx, you
> will find 100's
> of instances of ntwx being set to 50, 100, - small numbers
> that apparently
> don't make sense for most purposes. The more that we strive
> to make our
> stuff work well for our users, the more they will appreciate
> us, reference
> us in their publications, and keep coming back for more.
> Best Regards - Bob
>
> ----- Original Message -----
> From: "Adrian E. Roitberg" <roitberg.qtp.ufl.edu>
> To: <amber-developers.scripps.edu>
> Sent: Friday, May 05, 2006 9:58 AM
> Subject: Re: amber-developers: Troubles at PSC
>
>
> > Robert and list:
> > As for the guidelines, I always thought that you give the
> user enough rope
> > to
> > hang themselves with ;-) However, by reading the examples
> in the manual,
> > the test cases, and
> > the tutorials, the user should get a good idea of what
> types of 'flags' to
> > use.
> > Basically, the tutorials from Ross use reasonable time steps, etc.
> > if the user wants to try running amber without looking at
> those, adding
> > text into the manual warning them will not quite help either.
> > Cheers
> > Adrian Roitberg
> >
>
>
>
Received on Sun May 07 2006 - 06:07:06 PDT
