[jumbojacks]

I've never really seemed to understand how AMD/Intel are affected in the short term, but I think I have a much better feel of where both companies will be in the future.

I've always felt that AMD lacked in one area Intel seemed to have such a huge dominance over and that is providing platforms to go with the chips they produce (think chipsets and packages like the Centrino brand). Having "AMD endorsed" platforms could definitely increase AMD's acceptance among major OEM manufacturers and this is already being seen with Dell/HP. An area AMD seemed to have dominated Intel recently in terms of performance/power is in their server markets with their Opteron platforms where typically margins are much higher. The growth in marketshare in this segment has definitely hurt Intel.

In terms of pure performance, this is a period of computing history where I think you'll see the least amount of performance gains *in today's applications* due to the transition to multicore processors. It's basically the equivalent of a hack (can't improve the single core due to heat and other process limitations, so throw in more cores). There's a lot of emphasis on parallel computing that's starting to find it's way down at the undergrad level at major engineering colleges.

In the future, I think AMD and Intel will take very different approaches to tackle the problem of performance and parallelism. Their fundamental architectures are really starting to diverge where AMD is looking into specialized processors connected via HT links and Intel seems to just be throwing more cores at the problem but in an efficient design. I'm not really sure which implementation will succeed, but it should be interesting to see how things pan out in the future.