Expansion of Stam method into 3D

Martin Whyte

Hi, I've recently been looking at Jos Stams "Real-Time Fluid Dynamics for Games" paper (http://www.dgp.toronto.edu/people/st.../pdf/GDC03.pdf)

I've build the basic solver, and done some other bits, but now I'd like to try to expand it to three dimensions. This should be a straightforward exercise but I'm not sure where I've went wrong. How would I figure out what's went wrong? I think the problem lies in my interpolation. I expanded the 2d to 3d based on http://en.wikipedia.org/wiki/Trilinear_interpolation. Also with regards to the SetBnd() function, should I just be adding an extra term for when b==3, and also looping round twice instead of just once (since there's now a whole "plane" of wall values along the extents of XYZ, instead of just a row).

I've hosted the file here in the offchance someone wants to look at it (certainly I don't want this to turn into a "fix my code please" scenario, but I'd hate to be barking up the wrong tree entirely with my hypothesis of where the problem is. Also pasting large cuts of code in here seemed like a bad idea).