> Interesting - if you replace the sqrt(...) in the isosurface with
> sqrt(max(0,...)) it should work reliably with all versions (but is slow
> as hell for some reason).

In general, somewhere before version 3.6, isosurface{} and parametric{}
started taking MUCH longer to render than earlier.

As far as the sqrt of negative numbers is concerned ... I use Linux and Mac
OS-X and neither had problems, I suspect it it their use of the IEEE
libraies which hande those exceptions very well, I don't think MS/compilers
do so as well.

-------------------------------------
P a u l   B o u r k e
http://local.wasp.uwa.edu.au/~pbourke/

Post a reply to this message