So have we reached the top of the graph predicted by Moore's law? Seems like it (if not the top, very close to it), especially with respect to performance of a single, lonely processor. Herb Sutter has written a very nice article on how we are reaching the end of the road, or rather a bend in the road, which will require us to be much more aware of concurrency than we are at present. The article is slated to be printed in Dr. Dobb's Journal and the C/C++ Users Journal, in 2005 - but you can steal a sneak right now.
Interestingly, this article on the PC of 2005, in PCWorld, also mentions this exact same phenomenon, stating that few major hardware changes are expected in 2005. It even starts off with "Consumers thinking about buying a new computer in 2005 might be better off putting off their purchase until 2006" which is a very brave thing to say for a magazine which relies mainly on the health of the PC Industry!
So this might be a good time to read up on software concurrency and how it relates to development and testing of software. I know I need to.
P.S. - Wishing everyone a Happy New Year! May it be a peaceful and joyous one for all.