Lars Arge sent me a reminder, via John Hershberger, that the submission deadline for the Workshop on Massive Data Algorithmics, which will immediately follow SOCG 2009, is this Friday, April 3.
Meanwhile, Hughes Hoppe visited the CS department at UIUC today to give a nice colloquium talk about his recent work on some of his work on parallel spatial data structures.
Same thing, same thing. What kind of algorithms do you think are interesting for an architecture with 800 cores, hardware-level threading, but only a few kilobytes of cache for each cluster of 8 cores?!
Comments