Originally posted by: PhoenixOrion
Intel is no way close to being "screwed" just because cpu performance crown is lost back and forth one year to the next.
As long as Dell and other oem still puts corporate orders of 2000+ intel units here and there on strong marketing and label of "stability," that's a tough cookie jar to get amd hands into.
I'm just using my company for example on recent order of 2000 from Dell for two east coast sites.
Like I said, that is why the Opteron is so extremely dangerous. It gives AMD a small opening into the cookie jar that they can exploit if they play their cards right. I think for Dell to even think of picking them up, they would need a LOT more fabrication capacity than they already have. To ramp that up, they need serous income first.
Their strategy makes a lot of sense, when you think about it. They first attack the smaller markets: the server markets, the enthusiast market and the retail desktops. In these smaller markets they grow their marketshare and profits, then use this profit to fund more fabs in order to give them the capacity to keep growing their sales.
The big prize is of course Dell. They may buckle under and start selling Opteron servers if they start seeing the demand from customers, and AMD has the capcaity to handle taking on Dell in that smaller market. Another possibility is that if the high-end P4s stay stuck at their current performance level long enough, allowing AMD to get a pretty sizable lead in desktop performance, then Dell might start looking at the FX chips for their XPS series (also a smaller segment of Dell's business that AMD could likely handle). If either of these happen, it gives AMD a real chance to grow and become a much more dangerous competitor.
Not saying it WILL happen, just that it COULD, depending on how the next year or two plays out.
And as for Prescott being a smokescreen, I would have agreed if they had stuck to the old S478 for the Prescotts. The introduction of new DDR2/LGA775 Prescott-based boards make me doubt that they have anything else to pull out of their hat soon, because most of the new technologies in the 900-series boards (better able to handle higher voltages, support for more memory bandwidth) would benefit a high-clocked Prescott more than a Dothan-derived desktop.
In fact, why even bother to lengthen the pipelines of the P4 with the core redesign if there never was any intention of clocking the Prescotts up to the point where you would start to see a preformance increase over the Northwoods? If Prescott truly was a smokescreen, they could have just have just shrunk the Northwood architecture down to 0.09 micron and called it a day. A chip like that would have been even cheaper to fab because it wouldn't have needed 1MB of cache to make up for the loss in clock-per-clock performance of the longer Prescott pipeline. A smaller, cheaper core would have increased Intel's profits. If Prescott was a ruse, it was an expensive one.
No, I think never was any plans for anything besides a put out a line of 4-5GHz Prescotts that Intel would have ramped in just as quickly as the Northwoods in order to stay one step ahead of AMD. Something went terribly wrong with this plan (probably heat issues) and now they are most likely rushing out a desktop Dothan as a stop-gap, while at the same time slapping on a numbering system to make the changeover less apparent to the ignorant masses.