---------- Forwarded message ---------- From: John Peterson <jwpeter...@gmail.com> Date: Thu, Jul 14, 2016 at 8:52 AM Subject: Re: [Libmesh-users] Bug in PetscVector::create_subvector To: Jonas Ballani <jonas.ball...@epfl.ch>
On Thu, Jul 14, 2016 at 7:08 AM, Jonas Ballani <jonas.ball...@epfl.ch> wrote: > > Apparently, the vector y is not distributed correctly to the processors. Is > this a bug? Sure seems like one. PetscVector< Number > y( mesh.comm(), x->size() ); You are initializing the subvector "y" before passing it to create_subvector(). This constructor ends up calling: this->init(n, n, false, ptype); i.e. setting the global and local sizes equal to n. When the Vec type is AUTOMATIC and n==n_local, this ends up creating a SERIAL Vec under the hood, so I think this explains (first,last) being the same on both procs. There is a different code path in create_subvector() for when y is not already initialized... can you try that instead? -- John -- John ------------------------------------------------------------------------------ What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic patterns at an interface-level. Reveals which users, apps, and protocols are consuming the most bandwidth. Provides multi-vendor support for NetFlow, J-Flow, sFlow and other flows. Make informed decisions using capacity planning reports.http://sdm.link/zohodev2dev _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
