Paul,
Even if Hammer is fast out of the gate (in a year and a half), can AMD afford to keep developing a process roadmap, like the one Wanna BMW posted below, to keep improving it like they must to keep up with Intel?
intel.com
Roadmap for P4 gets it to 10 GHz, when, of course, one of those two digit nanometer processes is ready. AMD has nowhere near the money to spend anymore to get beyond 130 nanometer, which they haven't even achieved yet, much less 90, 65, 45 and 32 Nm, which will be needed to get up to that kind of clock speed. I pulled those from that article on Intel process development future plans. Those nanometer numbers are probably a complete don't care, No-op to AMD because where will they get the money to be able to afford to develop those capabilities? Just this year alone, Intel is spending $4.2B in R&D, about equal to AMD's run rate for revenues. After AMD's five consecutive losing quarters are up late next year, they won't have a pot to pi** in, much less the billions of dollar it's going to take to keep up. Hammer could be a footnote in microprocessor history by three or four years from now.
Tony |
