I rebuilt the PETSc/SLEPc and libMesh, but still fails to output .e file.
I am not clear the reasons, but it works well if I use GMVIO and VTKIO. The
ParaView can read the output files for these two cases. :-(
On Thu, Jul 30, 2015 at 1:45 PM, Xujun Zhao <xzha...@gmail.com> wrote:
> Yes, I enabled parmesh when configuring libmesh.
> Probably as you said, there is something wrong with my compiler/build. I
> will try to rebuild everything and see if it works.
> Thanks, John.
>
> Xujun
>
> On Thu, Jul 30, 2015 at 1:30 PM, John Peterson <jwpeter...@gmail.com>
> wrote:
>
>>
>>
>> On Thu, Jul 30, 2015 at 10:51 AM, Xujun Zhao <xzha...@gmail.com> wrote:
>>
>>> Hi all,
>>>
>>> Writing out equations systems failed during run, which used to work
>>> before.
>>> Just go to examples/system_equations/ex1/, make and run, it gives the
>>> following error msgs:
>>>
>>> --->test: write out the Exodus file ...
>>>
>>> Warning: This MeshOutput subclass only supports meshes which have been
>>> serialized!
>>>
>>> [0]PETSC ERROR:
>>>
>>> However, if I comment on the following line, everything is good.
>>> ExodusII_IO(mesh).write_equation_systems ("out.e", equation_systems);
>>>
>>
>> You configured libmesh with --enable-parmesh, right? I ran the same
>> example, swapping out Mesh for ParallelMesh and it works just fine.
>>
>> Furthermore, this example is run by the continuous integration system
>> every time someone opens a PR in libmesh, so I'm pretty confident it works.
>>
>> I suspect something is wrong with your compiler/build instead...
>>
>> --
>> John
>>
>
>
------------------------------------------------------------------------------
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users
