Hi, I am observing some really really really weird SNES behavior in my SNESVI
code called through Libmesh. This behaviour appeared after I changed some
initial conditions and only happens in an optimized build. I am running this
code on a Macbook pro running os x 10.8..
When the debug code is r
Hi,
This is very weird and I have no idea what has gone wrong. I am seeing no
output from my explicit system that I use to compute my results. I have no idea
what has gone wrong as this was working. I get fields of 0 instead. Does
anybody have any suggestions?Thanks, Subramanya.
Thanks for the warning. I'll let you know if I encounter any problems.
On Sat, Jul 13, 2013 at 4:31 PM, Derek Gaston wrote:
> Oops - I didn't read to the bottom where you said you got it
> solved I'm still wary of your MPI setup though...
>
> Derek
>
> Sent from my iPad
>
> On Jul 13, 2013,
Oops - I didn't read to the bottom where you said you got it
solved I'm still wary of your MPI setup though...
Derek
Sent from my iPad
On Jul 13, 2013, at 4:44 AM, Onur Solmaz wrote:
> Thank you for the detailed answer. I certainly did not have much idea on
> the strategies and paradigms p
Your problem is most likely that you're not using a consistent mpi.
--download-mpich is generally a bad idea with libMesh because it
installs a "private" mpi that is used to compile Petsc and if you
don't set your paths up to point to it when you compile libmesh then
you will be using two different
Thank you for the detailed answer. I certainly did not have much idea on
the strategies and paradigms put into creating interfaces and libraries.
As for my problem, I checked libpetsc.so found out that the symbols aren't
missing after all, so as you say it is a linking error. I compiled PETSc
simp
