SIAM is holding a conference on parallel computing in Seattle next February. I’m pretty psyched about it mostly because I’ve been sinking a lot of time into GPU-based stream computing lately and from a cursory look at the program I’d say that GPGPU is going to be the subject of a lot of interest.
For example here is an abstract from one of the Joint JSIAM — SIAM Lectures (sortof amusingly entitled: GPU Acceleration: A Fad or the Yellow Brick Road onto Exascale?):
“Since the first commodity x86 cluster Wigraf achieving paltry 10s~100s Megaflops in 1994, we have experienced several orders of magnitude boost in performance. However, the first Petaflop was achieved with the LANL RoadRunner, a Cell-based “accelerated” cluster, and in 2010 we may see the first (GP)GPU-based cluster reaching Petaflops. Do such non-x86“accelerator” merely push the flops superficially, or are they fundamental to scaling? Based on experiences from TSUBAME, the first GPU-accelerated cluster on the Top500, we show that GPUs not only achieve higher performance but also better scaling, and in fact their true nature as multithreaded massively-parallel vector processor would be fundamental for Exascale. Such results are being reflected onto the design of TSUBAME2.0 and its successors.”
In general the computer hardware industry is unpredictable and early adopters often get burned. How widespread GPGPU eventually becomes still remains to be seen, but such talk must be encouraging to anyone who has invested time into constructing and implementing algorithms on GPU architectures. Anyways, I’m hoping to make the trip up and do a poster presentation of some of the GPU-based matrix compression work I’ve been doing.
January 25th, 2010 - 19:32
Aw, this was a really quality post. In theory I’d like to write like this too – taking time and real effort to make a good article… but what can I say… I procrastinate alot and never seem to get something done.