Closed kswartz92 closed 2 weeks ago
I was looking through the physics modules and noticed that ode_time_point_
isn't updated in reverseAdjointTimestep()
. The check-pointed time steps and states are grabbed, but the absolute time doesn't seem to be decreasing during the reverse pass. Could this be the issue?
I have a branch that addresses all of this. I need to run your problem to make sure the dirichlet time dependence works now.
Fixed by #1183
@tupek2 I located a very small change (adding time to essential bc) that breaks the sensitivity test, here is a branch with the change https://github.com/LLNL/serac/tree/swartz10/transientEssentialBCtest