Charles - <I do not seem to recollect if PB's claims were limited to announcement or if he meant real availability. >
I didn't specify. If people are thinking about my bet with kash wrt to my comments, that's for the end of the year and Willamette vs. AMD whatever-core. As for Athlon vs. Cumine in the nearer term, my WC for Intel was single digit weeks behind. I still hold out hope (given my bias) that that WC still holds.
I've said it will be close to 1GHz. However, I've repeatedly made pains to acknowledge that the Athlon has the advantage wrt to design scalability, and it very well might win.
Both are limited availability now. However, if AMD ramps to the schedule they publicly committed to, and Intel does not release commercial quantities until Q3, AMD will obviously be ahead in terms of raw clock speed.
If Intel steps up its efforts to yet again try to accelerate it's MHz ramp remains to be seen. But based on what Intel and AMD is publicly saying about availability, AMD would have the advantage with the highest speed grades for the next 1-2 Q's. Beyond that, I would not want to hazard a guess at this point, other than my bet with kash, and I still feel pretty good about that one.
BTW, I do not see this situation as part of poor yields or any such thing on the part of Intel. Actually, they are quite good, even at 1GHz. Intel probably could do some things that could potentially increase the odds of better availability at the top speed grade, but there would probably have to be a trade-off wrt to capacity to run an accelerated schedule to accomplish this.
I give AMD credit for the Athlon design's scalability, and ramping their .18 process in a stable fashion without significant manufacturing issues. AMD has executed very well to date. Intel will have to tough it out until Willamette, keeping the gap as narrow as possible, and maybe (a suggestion here) concentrate on improving supply at a spot a speed grade or two below, i.e., huge quantities. I believe Intel could do this as capacity increases over time.
PB |
