HdivB/B rapidly increasing in colliding flows

[teogeo1996]

Whilst running simulations including the creation of a shock compressed layer, the value of HdivB/B rockets well above one even after setting hdivbbmax_max=512 as recommended by the code. Changing overcleanfac to account for this essentially makes the code extremely slow. Is there another way to go about this issue?

[danieljprice]

not really, at some point you have to pay the cost to get div B lower... the only other thing here is that we use a cleaning speed of sqrt(cs2 + va2) so if you're highly super-Alfvenic then a higher cleaning speed (overcleanfac) could help without killing the timestep too much...

[danieljprice]

we recently pushed a fix that may fix the problem here, see #537