Okay, I see your point about this being wierd behaviour for ntx > 3, irest
0. So the most simple solution is to disallow this combo - error exit when
you hit it. Then fully doc the combinations of ntx, irest that ARE
permitted, and what they do. We have full functionality already, but we
have a buried implied meaning in the ntx value. I actually see no real
problem with making nxt > 3, irest 0 work either, as it might be expected;
we just have to grind through all the combos that are possible (really only
4) and make sure that all the ones that currently work to do something
sensible continue to do the same.
So quick rundown:
ntx < 3, irest 0 - currently makes sense (no restart, set random
velocities) - supported, keep
ntx < 3, irest 1 - currently makes no sense and is currently thrown out -
error check supported, keep
ntx > 3, irest 0 - currently does the wrong thing, make it work in future
(no restart, set random velocities, since this is doable)
ntx > 3, irest 1 - currently makes sense (restart) - keep
You all let me know which way you go with sander, and I will do the same in
pmemd (options - 1) to either add error check for ntx >3, irest 0, or 2)
make ntx > 3, irest 0 set random velocities after zeroing the read
velocities (you pretty much have to read the vels to get to the box anyway,
and the setup work is negligible).
Best Regards - Bob
----- Original Message -----
From: "David A. Case" <case.biomaps.rutgers.edu>
To: "AMBER Developers Mailing List" <amber-developers.ambermd.org>
Sent: Wednesday, January 07, 2009 4:35 PM
Subject: Re: [AMBER-Developers] inconsistency in sander manual and code
forrestart velocities
> On Tue, Jan 06, 2009, Carlos Simmerling wrote:
>
>> hi all- I've been tracking down a bug in sander for amber9/10 and know
>> that one exists, but it's hard to say if it's in the manual or the code.
>> I just want some quick feedback since there could be several ways to
>> "fix" it. it's related to all of the overly complicated combinations of
>> irest, ntx, tempi and init.
>>
>> my understanding, along with page 26 of the amber10 manual, is that
>> for a PBC system, if one wants to restart with box info from inpcrd
>> (say from a constant pressure run), then one should use ntx=5.
>>
>> from the manual:
>> ntx = 5 X and V are read formatted; box information will be read if
>> ntb>0. The velocity information will only be used if irest=1.
>
> I need to check for sure, but I *think* the following is a correct
> statement about sander:
>
> Box information is always taken from inpcrd, at least for formatted
> restarts. If this is broken for unformatted restarts, that is a bug.
> [This means that the box is read from inpcrd even if ntx=1. Hence the
> workaround for Carlos' situation would be to use ntx=1.]
>
>>
>> it's the last part that's wrong. if I set ntx=5 to read the box, but
>> irest=0 so that my velocities are assigned by sander, setvel() is
>> still not called in sander.f (it is only called if ntx <=3). this
>> means the run is the same no matter what the random seed is. also,
>> init is passed to setvel but never used.
>
> I appreciate the desire not to break backwards compatibility, but the
> behavior described here really seems like nothing but a bug. If
> irest=0, the velocities in inpcrd should be ignored, and setvel() should
> be called to assign velocities.
>
>>
>> SO, I think the problem is that we use ntx for deciding on calling
>> setvel, but we really should use ntx for reading them, and irest for
>> calling setvel...
>
> I'd appreciate detailed proposals, but right now, I don't see anyway to
> have reasonable behavior without at least a small break from the past.
> I doubt (hope?) that there aren't many scripts out there with ntx=5 and
> irest=0. (People might poke around their directories to check this.)
> And, of course, we need to fix the documentation about what happens when
> ntx=1.
>
> ...dac
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Jan 09 2009 - 01:13:41 PST
