“A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.”
― Douglas Adams, Mostly Harmless
One could possibly add a comment to test input files like “For testing purposes only…”
:-)
Gustavo Seabra
> Em 18/08/2015, à(s) 16:50, Adrian Roitberg <roitberg.ufl.edu> escreveu:
>
> I agree.
>
> But then the test cases will have a fixed ig ? I am all for it, but....
>
> I am very afraid of people lifting the mdins from the test cases and
> forgetting to change ig...
>
> I guess it is a case of foolproofing creating more fools ?
>
> adrian
>
>
> On 8/18/15 3:48 PM, Hai Nguyen wrote:
>> +1 for default ig=-1. I myself got periodic result a while ago and need to
>> rerun my sims.
>>
>> Hai
>>
>> On Tue, Aug 18, 2015 at 12:13 PM, Jason Swails <jason.swails.gmail.com>
>> wrote:
>>
>>> -------- Forwarded Message --------
>>>> From: Adrian Roitberg <roitberg.ufl.edu>
>>>> Reply-to: AMBER Mailing List <amber.ambermd.org>
>>>> To: AMBER Mailing List <amber.ambermd.org>
>>>> Subject: Re: [AMBER] NPT equilibration failed
>>>> Date: Tue, 18 Aug 2015 09:02:38 -0400
>>>> Also PLEASE PLEASE PLEASE add ig=-1 to your langevin thermostat inputs!
>>>> Oterwise you can see unphysical behaviors. Please look at the manual for
>>>> the origin of this requirement.
>>> I think we should change the default value of ig to -1. Keeping the
>>> default value is almost always the wrong thing to do, but it is not
>>> obviously wrong, which makes it a dangerous default in my opinion. I
>>> think that ig=-1 is almost always a better default for ig and never a
>>> worse one.
>>>
>>> The main issue with changing the default is going to be updating the
>>> tests to specify an explicit seed where the current default is relied
>>> upon for reproducibility. But that is easily worth fixing the default.
>>>
>>> Does anybody have particularly strong objections to this?
>>>
>>> Thanks,
>>> Jason
>>>
>>> --
>>> Jason M. Swails
>>> BioMaPS,
>>> Rutgers University
>>> Postdoctoral Researcher
>>>
>>>
>>> _______________________________________________
>>> 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
>
> --
> Dr. Adrian E. Roitberg
> Professor.
> Department of Chemistry
> University of Florida
> roitberg.ufl.edu
> 352-392-6972
>
>
> _______________________________________________
> 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 Aug 21 2015 - 14:30:03 PDT
