tap tap tap tap tap tap....

blah blah blah
Filling in that time while waiting for the currently running model number crunching to pass from "reasonable duration" (about 10 min), to "I'm not actually running but eating my own tail and disappearing up my own arse thank you very much". It would be nice to get results out the former, but the latter is becoming increasingly more likely.

I could tell that the last stellar model wasn't quite right. It didn't have any light coming out, somewhat contrary to the observations...

Round 1 suspected problem:
Input model not large enough so shock falls at boundary, gets munted and everything goes down the tube with NaNs spilling out left right and centre.
Solution 1:
Increase layers further out. Cut and paste in more with only tau increasing.
Result 1:
Eating tail territory.
Solution 2: (after stopping tail munching)
Increase layers further out but remember to change radius as well as other parameters...
Result 2:
...

Bored yet? I know I am...
5 min to go...
And pocket Oxford beats LJ spell checker any day.

And lets crash that - Oh look, some new error messages - wonder what those mean?
ATTENTION...KAPPA-TAU-INTERPOLATION POOR IN RING 32 32

[Edit] Translation: opacity (how mirky things are) optical depth (how far you can see in) interpolation mucking up in these layers...maybe

Solution and results 3 at a later date. [Edit] One additional step being tried. Being left running after head for bus I suspect...