"AMD executives have quietly begun highlighting the risks involved with shifting architectures so quickly."
Do AMD have a point? Is 32nm really a neccesary advancement, or are they simply sore at the prospect of being so far left behind.
Having an AMD rig, i am excited by Phenom II, but also wondering how long AMD will be able to realistically compete, considering that Intel's 32nm is progressing so well.
Guess we will find out more about 32nm next week.
AMD has 32nm samples out from Fab36 in Dresden and is slated production release for 2010.
Of course AMD will continue and be able to compete, if they haven't gone bankrupt or been bought by now with everything that's happened over the past three years, they're not going anywhere anytime soon.
Not sure what the risks are as the result of shifting architectures so quickly, aside from saturating the market and forcefully retiring parts before the market has had a chance to adopt and grasp the technology.
Is 32nm a "necesary" advancement? No, it is not. Fact is, for the average enthusiast and office worker, quad core 45nm is more than enough. As said time again, software and I/O need to catch up before we can truly take advantage of all those cores.
IMO, the only interesting result from 32nm would be the combo CPU/GPU system on a die type packages. Now that x86 code is being written for MID's, UMPC's, smart phones, PDA's, and micro-PC type devices, the combo CPU/GPU would be a great step forward in mobile computing.
Whether AMD or Intel, I hate the prospect of a new chipset everytime a new line of processors is released. Part of the reason Skt939, AM2, and Skt775 has done so well for AMD and Intel was the simple fact that we could simply drop the proc in a update the BIOS to realize the performance gains of new tech.