About 2 ½ years ago I wrote a fairly pessimistic post about the utility of multicore chips. It’s here. I read this morning in the July issue of IEEE Spectrum an article by David Patterson, a famed computer researcher at UC Berkeley, on the same subject. His conclusions are approximately the same as mine, but he states them more eloquently, in greater detail, and with far more credibility. He also offers some predictions, and a few rays of hope.
Patterson’s article is available online here. I recommend it.