- changed status to wontfix
adaptive solver demo (auto-adaptive-poisson) fails to run in parallel
Issue #985
wontfix
The auto-adaptive-poisson demo fails to run in parallel, giving the error message:
Traceback (most recent call last):
File "demo_auto-adaptive-poisson.py", line 253, in <module>
solver.solve(tol)
File "/usr/lib/python3/dist-packages/dolfin/fem/adaptivesolving.py", line 81, in solve
cpp.AdaptiveLinearVariationalSolver.solve(self, tol)
RuntimeError:
*** -------------------------------------------------------------------------
*** DOLFIN encountered an error. If you are not able to resolve this issue
*** using the information listed below, you can ask for help at
***
*** fenics-support@googlegroups.com
***
*** Remember to include the error message listed below and, if possible,
*** include a *minimal* running example to reproduce the error.
***
*** -------------------------------------------------------------------------
*** Error: Unable to perform operation in parallel.
*** Reason: Extrapolation of functions is not yet working in parallel.
*** Consider filing a bug report at https://bitbucket.org/fenics-project/dolfin/issues.
*** Where: This error was encountered inside log.cpp.
*** Process: 0
***
*** DOLFIN version: 2017.1.0
*** Git changeset: unknown
*** -------------------------------------------------------------------------
The first iteration finishes successfully, the error occurs when preparing for the second adaptive iteration.
Comments (3)
-
-
reporter Thanks for the confirmation of status. Good fodder for a PhD project.
-
How big of a project are we looking at here? Are we just missing parallel extrapolation of functions? What in particular makes this more difficult to implement in parallel than other FEniCS features?
I have been wanting to dig into this question for a while now. Is there already a nice summary anywhere that would help me better understand it? Or do I just start by looking at the code?
- Log in to comment
This is known not to work in parallel. Any suggestions for a suitable fix would be welcome.