Some fun…
Something I’ve been meaning to play with for a while, inspired by a slashdot post the other day:
Hopefully I’ll be able to speed up the calculations enough to have it work on more than just .1% of the data in reasonable time (the above took three hours of CPU time to generate, sadly), at which point we might be getting somewhere.
Are you interested in using NVIDIA’s CUDA or the only-just-released OpenCL?
With the power of a GPU or two you could speed this up enormously.
I am looking forward to OpenCL, which should be released publicly by NVIDIA in a couple of weeks judging by phoronix.com stories.