Progress in ray tracing
Way back in 98 or so, Chris Woody and I built a 8-CPU Linux cluster using NFS, Povray and homebuilt task partitioning code in C. Ahh, good times - thinnet & hand-installs on freaky flaky hardware.

Overall, we got about half the performance of a Pentium-II, which was decent, mostly due to the efficiency with which Povray can slice up a rendering job. You can see more details on the project page.
I bring this up because today I see the newest nVidia hardware/software that does raytracing a la Povray… in real time, at 30 frames per second, at 1920×1200:
Yowza. The times, they are amazing.

