[ESPResSo-devel] FW: Local stress tensor/Multibody interaction and A bug in tensor distribute

[Zunjing Wang]

From: Zunjing Wang <address@hidden>
Date: Tue, 16 Jun 2009 20:48:26 -0400
To: Christoph Junghans <address@hidden>
Conversation: [ESPResSo-devel] Local stress tensor/Multibody interaction and
A bug in tensor distribute
Subject: Re: [ESPResSo-devel] Local stress tensor/Multibody interaction and
A bug in tensor distribute

Dear all,

Thanks for all the discussions here.

On 6/16/09 4:57 PM, "Christoph Junghans" <address@hidden> wrote:

> Hi all again,
> 
>> The pressure might not depend on them globally (but I'm not fully
>> certain here), but it surely matters locally.  So whenever one wants a
>> spatially resolved stress tensor, these terms matter.  Since many
>> people now use ESPResSo with more than just two-body-potentials, none
>> of them can use the local stress tensor.  Are there efforts on the way
>> to fix this?  (No point to start doing this when somebody else has
>> already done so).
> You are totally right it matters for any local pressure quantity and
> it also matters for the stress tensor. Only the total pressure (a
> number) is unchanged, (at which I am looking most of the time). I did
> not remember the reference now, but I guess it was this one:
> http://pubs.acs.org/doi/abs/10.1021/jp952646x
> 
> I think it is not too much effort to change it. add_bonded_virials is
> the function to change.
> 

>>> It will depend very much on the system you study how reliable the
>>> results are without these contributions.
>> 
>> Strange answer... If you're sure that the multibody terms don't affect
>> the global pressure, then the answer should be that there will be no
>> effect whatsoever from leaving them out.  Maybe you're not sure?  For
>> the local stresses your answer might be true, though.
> You are right, I did not make my point very clear. It matters for the
> stress tensor in any case, but the question is how big are the forces
> from the angle potential in comparison to the 2-body forces. Maybe
> this contribution is just small and negligible (but this will depend
> on the system studied).
> 
>> Sure it's a feature: It's a security check that tests whether the
>> distribution of the virial between the boxes is done completely
>> (without bits missing).  In our case the error can become as big as
>> 1e-8, which is way too large, a million times bigger than
>> ROUND_ERROR_PREC.  Something clearly seems to be wrong here, yet I
>> would not know why this should depend on the system.  Seems like we
>> need to go through the local stress tensor routine and check it...
> I guess in the worst case it could be:
> (number of addend in loop)*ROUND_ERROR_PREC
> We should take a look at size of the boxes. Btw. why is not
> fabs(length-calclength) in condition of the error message.

In my opinion, it should also be fabs(length-calclength) rather than
length-calclength. I tested it again today, finding that the (number of
addend in loop) [which is only in the order of magnitude ~10] is independent
of the fabs(length-calclength) [which varies at least through over 1E-16 -->
-1E-8/even uper]. Sounds to me there is a problem in distributing the virial
in the current Espresso.

> 
> Cheers,
> 
> Christoph
> 
> 
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
> 
> 
> 
> _______________________________________________
> ESPResSo-devel mailing list
> address@hidden
> https://fias.uni-frankfurt.de/mailman/listinfo/espresso-devel
>