On Thu, 28 Jan 2016, Junchao Zhang wrote:
Roy, I tested your patch and found it did not fix the problem.
Oh, no, it wasn't expected to fix your problem, I'm afraid - it was
just step 1 of the fix for your problem. Paul Bauman is promising to
ride to our rescue on steps 2+; if he doesn't fi
Roy,
I tested your patch and found it did not fix the problem.
I can send you the modified example if need.
--Junchao Zhang
On Thu, Jan 28, 2016 at 2:25 PM, Roy Stogner
wrote:
>
> On Wed, 27 Jan 2016, Roy Stogner wrote:
>
> 1. DofMap::process_constraints() needs to be fixed to properly upda
On Wed, 27 Jan 2016, Roy Stogner wrote:
> 1. DofMap::process_constraints() needs to be fixed to properly update
> adjoint and not just primal rhs'es when expanding constraint
> equations. I'm doing this myself right now, partly because I know
> even more people it could affect, partly because it
This is my workaround. I dialed the clock back and forth around
equation_systems.reinit().
I am not sure if there is a more effective way (in performance) to do so.
system.time = 0;
Create a DirichletBoundary object and add it to equation_system;
equation_systems.init ();
for each time step {
s
On Wed, 27 Jan 2016, Junchao Zhang wrote:
> By the way, in the last few hours, I debugged the code and found these
> comments in TransientSystem::reinit (). I traced through the class
> hierarchy from TransientSystem, LinearImplicitSystem, ImplicitSystem,
> ExplicitSystem to System, but found
On Wed, 27 Jan 2016, Roy Stogner wrote:
> That's nothing I'd be able to get to myself in the immediate future,
> although obviously "subtle inaccuracy when mixing time-dependent
> Dirichlet + AMR" ought to be a high priority to fix.
On the other hand, Paul Bauman just IM'ed me to say that he'd b
Thanks for the info. Let me try the workaround.
By the way, in the last few hours, I debugged the code and found these
comments in TransientSystem::reinit (). I traced through the class
hierarchy from TransientSystem, LinearImplicitSystem, ImplicitSystem,
ExplicitSystem to System, but found nothing
On Wed, 27 Jan 2016, Junchao Zhang wrote:
> I slightly modified adaptivity_ex2.C ( see the outline below). I find
> after the mesh is refined, the old solution is not correctly projected to
> the new mesh.
> I print out old solution in each refinement step. At the beginning, time
> = 0.025, r_s
