[Nek5000-users] Failure when changing geometry in usrdat2

[nek5000-users at lists.mcs.anl.gov]

Just a short update. I have tried the following:
-using PN-PN instead of PN-PN-2
-characteristic time stepping
-fiddling with the projection parameters P094,P095
-turning off the convective forcing (Rayleigh number 0)
without any change.
I also tried deforming the grid in the rayleigh example, which is a 
somewhat strongly simplified 2D version of my problem, and this worked 
perfectly.

Jan