[aspect-devel] Checkerboard at large scale

[Wolfgang Bangerth]

On 04/11/2018 07:04 PM, Lev Karatun wrote:
> 
> just wanted to let you know, it seems I did have a slight imbalance in the 
> boundary velocity, and after fixing it (and making net volume flux equal to 
> zero) the solver successfully converged with constant mesh. It still chokes 
> when I enable AMR, so I'll keep the free surface enabled as you suggested.

Nice! It's good to know that our collective hunch was correct.

Solving these sorts of issues is hard. It requires intuition for where to look 
and what the problem could be. It's good to know that collectively, we have 
this sort of intuition. It's also good to know that we can have trust in the 
correctness of the algorithms implemented and do not always have to assume 
first that there is a bug in the code.

Thanks for walking through this process with us!
W.

-- 
------------------------------------------------------------------------
Wolfgang Bangerth          email:                 bangerth at colostate.edu
                            www: http://www.math.colostate.edu/~bangerth/