Monday, February 27, 2006

IBM's Octopiler, or, why the PS3 is running late

by Hannibal

All afternoon I've been slogging through IBM's 25-page paper on their newly released Octopiler, and now things are clearer to me. See, Cell's greatest strength is that there's a lot of hardware on that chip. And Cell's greatest weakness is that there's a lot of hardware on that chip. So Cell has immense performance potential, but if you want to make it programable by mere mortals then you need a compiler that can ingest code written in a high-level language and produce optimized binaries that fit not just a programming model or a microarchitecture, but an entire multiprocessor system. This isn't just a tall order, or even a doctoral dissertation. It's a generation's worth of doctoral research. Meanwhile, the PS3 is due out in 2006.
Octopiler is intended to become just such a compiler�one that can take in a sequential program that's written to a unified memory model, and output binaries that make efficient use of the massive, heterogeneous system-on-a-chip that is the Cell Broadband Engine. I say 'intended to become,' because judging from the paper the guys at IBM are still in the early stages of taming this many-headed beast. This is by no means meant to disparage all the IBM researchers who have done yeoman's work in their practically single-handed attempts to move the entire field of computer science forward by a quantum leap. No, the Octopiler paper is full of innovative ideas to be fleshed out at a further date, results that are 'promising,' avenues to be explored, and overarching approaches that seem likely to bear fruit eventually. But meanwhile, the PS3 is still due out in 2006

No comments:

Post a Comment