Re: [ESPResSo-users] Limited number of virtual particles

[Owen Hickey]

Hey Aleksei,

The issue is not actually virtual sites but the analysis routines. The
specific relate to memcpy versus
memmove:http://lists.gnu.org/archive/html/espressomd-devel/2015-01/msg00012.html

The following code should work now and soon be in the master:
https://github.com/ohickey/espresso/tree/fixMemCpy

For calculating the structure factor I would actually recommend
dumping out blockfiles of all the particle coordinates and writing
your own analysis routines in say python. Especially since you are
dealing with polymers and a static property, you should get decent
results without having to write out the particle coordinates very
often.

Happy simulating :)

Owen

On Thu, Jun 4, 2015 at 1:29 PM, Aleksei Kabedev
<address@hidden> wrote:
> Dear all,
>
> I need to implement a system with A-B block-copolymer chains and to put the
> virtual particles in the centers of mass of each block. Say, we have 2
> chains, consisting of 10 monomers each, where for both chains A part is a
> set of first 5 monomers (of type 1) and B part is second 5 (type 2). Then I
> put the virtual particles in the COMs of each set of 5 particles (type 3 for
> the com of 1 and type 4 for the com of 2). After that I study structure
> factor and RDF of particles of types 3 and 4. It works pretty fine with up
> to 15 chains, no matter how long the chains are. But if I use 16 chains or
> more it returns segfault. Is there any limitation related to the number of
> virtual particles that can be used or anything like that?
>
> By the way, probably it is another question to ask - is there any better way
> to get a structure factor of diblock-copolymers system using Espresso, than
> that?
>
> Please see the example of the code in attachments. My apologies for the
> mess!
>
> Thank you in advance!
>
> --
>
> Kind regards,
> Aleksei Kabedev