I’ve previously posted on the topics of CUDA and Larrabee. I continue to be intrigued by the possibilities that open up as multi-core GPU programming becomes available. For applications that need many threads this should present interesting opportunities. Why bother struggling to run your parallel application in the meager 4 or 8 cores of your CPU when you can offload the work to 32 cores?
In fact, with so many takes on this – it would only be a matter of time before we would see calls for consolidation of efforts. The latest issue (March ’09) of MaximumPC does exactly that. I would link to the article, but it isn’t online yet. I’m sure they aren’t the first to call for this, but I can’t say that I follow the graphics card industry very closely. In fact, MaximumPC’s call is as much a warning to NVidia made by drawing parallels to 3dfx and Glide. It’s a reasonable comparison. (And I remember that era…I bought a 3dfx card and played my share of GLQuake.) But the industry should have learned by now that a standard architecture and API will be more likely to succeed.
As I’ve said before, I’m looking forward to seeing how many cores we can throw at modern parallel rule engines – along the lines of Dr. Forgy’s recent efforts.