Michael Tupek
Michael Tupek
also correctly label const-ness of these functions which do modify non-mutable internal states.
Add a recursive call to advanceStep() to handle cases where the solver fails due to the boundary conditions changing rapidly. Also, add utilities to check if a matrix is symmetric...
Abstract solid mechanics a bit to allow for contact jacobian corrections in the adjoint calculations. Gradient test is not super accurate, maybe ~20% off. Hopefully the analytic contact jacobians can...
mpirun -n 8 ./tests/solid_reaction_adjoint seems to result in a segfault.
Add implicit (differentiable) linear and nonlinear solvers. Add time discretization for second order systems. Add the ability to compute resultant forces. Dynamic patch test and design sensitivity tests are included.
This makes it a bit hard to use, as users are wanting to pass parameters back in using the order specified in the functional constructor argument order. We may just...
[INFO] Saving data collection at time: '1' and cycle: '1' to path... This it output on every rank and fills up screens unnecessarily.
Seems like something easy to check at either runtime or compile time. This cost us some time debugging what should have been a non-issue.
The first iteration of this test does not converge. Any change in solver behavior breaks this test. Is this model still needed, is it well formulated (I suspect nonsmoothness, but...