[Gfs-devel] Patch for a bug in gfsvariablePoisson
54 views
Skip to first unread message
Jose M. López-Herrera Sánchez
Jan 12, 2014, 10:53:00 AM1/12/14
to GFS developper discussion list
Hi all,
Actually, the gfsvariablePoisson does not solve properly when solids are present. This is due to a wrong argument at gfs_poisson_coefficients (). The attached patch correct this issue.
Cheers
Jose
Vladimir Kolobov
Jan 5, 2015, 3:35:30 PM1/5/15
to gfs-...@lists.sourceforge.net
Jose,
In addition to your patch, one needs to add:
gfs_all_reduce (domain, v->centered, MPI_INT, MPI_MIN);
after "has_dirichlet" call.
This will properly set the "v->centered" flag for parallel runs. Otherwise, some procs (which do not have solid objects) will have v->centered = TRUE and this will mess up the things (especially the linear/hypre solver stencils)
Regards,
Vladimir
In addition to your patch, one needs to add:
gfs_all_reduce (domain, v->centered, MPI_INT, MPI_MIN);
after "has_dirichlet" call.
This will properly set the "v->centered" flag for parallel runs. Otherwise, some procs (which do not have solid objects) will have v->centered = TRUE and this will mess up the things (especially the linear/hypre solver stencils)
Regards,
Vladimir
------------------------------------------------------------------------------ CenturyLink Cloud: The Leader in Enterprise Cloud Services. Learn Why More Businesses Are Choosing CenturyLink Cloud For Critical Workloads, Development Environments & Everything In Between. Get a Quote or Start a Free Trial Today. http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
_______________________________________________ Gfs-devel mailing list Gfs-...@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/gfs-devel
Reply all
Reply to author
Forward
0 new messages