Can the 360 really live that long!

[bdbdbd]

@Mifely: So, what's stopping M$ to go to architechture that works like the Cell?
The genious of the Cell is, that every SPE works from its own cache, while PPE can work as a controller for all the SPE:s. The problem in the other hand is, that one SPE can't do much per cycle.
Sure, it's scalable architecture, but then again, adding the number of SPE:s requires more space or a size shrink (and shrinking the PPE leads us back where we started).
PPE is the "weak spot" in Cell developement, since shrinking it causes loss in number of SPE:s and if you don't shrink it, it causes loss in performance. So you aim for the smaller scale manufacturing tech, shrink SPE:s and add transistors for the PPE in order to use the space, when the biggest improvement was the huge cache Cell has. What happens, is that you have the same issues that you have with other processors too. Basically Cell developement goes forward to keep up the pace with competion, not that it would be some sort of miraclous supercomputer. The biggest advantage Cell has over the multicore (CISC) processors, is its price, not performance. And yes, Cell type of solutions may very well come more common in RISC processors, before the quantum processors hit the market.

Main reason for multiple processors is, that the "arms race" runs faster than the tech. Performance has to double faster than every 18 months. And this is just caused by HW manufacturers trying to show who's got the best performance. A little similar thing as during the cold war between USA and Soviet Union. Both had enough nukes to destroy all (not counting in bacteria) life on earth multiple times, but still had to race who could do it more often. So it's not the markets demands, it's the competitions demands.

The reason for jumping into next gen isn't the technical requirements, or the markets demand, but a financial decision of the manufacturers. You can't sell the same stuff for the same customers too many times and in order to raise the sales, when market starts to saturate, you need to give something new for the market. I believe that Sony and M$ are going to profit from hardware in the next gen, which means that early console release serves them both, in terms of getting a foothold from the market and still making profit, so the 2011 for 720 sounds very logical.

@Squiliam: X86 processors for the consoles doesn't sound propable, since the CISC processors tend to be more expensive than RISC.

And the Xbox was killed because M$ couldn't afford to make heavy losses with two consoles simultanously. The bad NVidia deal sure contributed a lot, but the decision was purely financial. They could have made a new deal, but supporting Xbox didn't make financial sense.