LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] reverse_comm_compute
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lammps-users] reverse_comm_compute

From: Steve Plimpton <sjplimp@...24...>
Date: Thu, 10 May 2018 07:01:53 -0600

yes.  There are many examples of fixes and computes that use
reverse communication.  You can look at those as examples.


On Wed, May 2, 2018 at 7:49 PM, Cunzhi Zhang <zcz1005@...2045...> wrote:
Dear lammps community

I am reading compute_stress_atom.cpp. there are some data communication function and parameters that confuse me.

pack_reverse_comm(int n, int first, int *buf): my understanding is all these parameter are automatically dealt by lammps. n: the number of nghost atoms with data need to be reduced; first: kind of like the total number of local atom, after which are index for nghost atoms; am i right?

Also, I should set comm_reverse to be the size of each atomic data (vector of 3 or tensor of 9, etc.)?

So, basically if I want to reduce my data, all I need is to reset comm_reverse, and modify the data transfer part in pack_reverse_comm and unpack_reverse_comm function?



Check out the vibrant tech community on one of the world's most
engaging tech sites,!
lammps-users mailing list