This is a read only copy of the old FEniCS QA forum. Please visit the new QA forum to ask questions

Unable to solve optimization problem (segmentation fault)

+1 vote

I am testing the topology optimization example possion-topology.py in the dolfin-adjoint source tree. Every time I ran it ended with "Segmentation fault (core dumped)". Even though I decreased the mesh size from 250 to 10, it also gave the same error. This is how it looks like:

******************************************************************************
This program contains Ipopt, a library for large-scale nonlinear optimization.
 Ipopt is released as open source code under the Eclipse Public License (EPL).
         For more information visit http://projects.coin-or.org/Ipopt
******************************************************************************

This is Ipopt version 3.12.5, running with linear solver mumps.
NOTE: Other linear solvers might be more efficient (see Ipopt documentation).

Number of nonzeros in equality constraint Jacobian...:        0
Number of nonzeros in inequality constraint Jacobian.:      121
Number of nonzeros in Lagrangian Hessian.............:        0

Hessian approximation will be done in the space of all 121 x variables.

Scaling parameter for objective function = 1.000000e+00
objective scaling factor = 1
No x scaling provided
No c scaling provided
No d scaling provided
Moved initial values of x sufficiently inside the bounds.
Current control integral:  0.0
Initial values of s sufficiently inside the bounds.
*** -------------------------------------------------------------------------
*** Warning: Function copy constructor has been deprecated in FEniCS version 2016.1.
*** It will (likely) be removed in the next FEniCS release.
*** Use 'Function.copy(deepcopy=True)' for copying.
*** -------------------------------------------------------------------------

Solving nonlinear variational problem.
  Newton iteration 0: r (abs) = 9.248e-04 (tol = 1.000e-07) r (rel) = 1.000e+00 (tol = 1.000e-09)
  Newton iteration 1: r (abs) = 1.043e-17 (tol = 1.000e-07) r (rel) = 1.128e-14 (tol = 1.000e-09)
  Newton solver finished in 1 iterations and 1 linear solver iterations.
Segmentation fault (core dumped)

Anybody can help? Thanks in advance. By the way, Ipopt has been successfully installed because I verified the test examples.

Update:

All topology cases end with segmentation fault but there are plenty memory left, memory usage is only 40%.

Li

asked Jul 18, 2016 by lz2m12 FEniCS Novice (290 points)
edited Jul 19, 2016 by lz2m12

the exactly the same script failed as Segmentation fault (core dumped) in Docker enviornment but success in virtualmachine. i am totally lost.

1 Answer

+1 vote
 
Best answer

Yes, I found the problem.Actually it's not a bug of fenics or dolfin-adjoint. the reason of the segmentation fault is something with MUMPS. It works in the fenics virtual machine but not in my computer. After I replace the linear solver with HSL, it works finally.

answered Jul 20, 2016 by lz2m12 FEniCS Novice (290 points)
selected Jul 21, 2016 by lz2m12
...