Re: it sounds like AMD delivered once
Boy, are you ever naive when it comes to Intel vs. AMD.
Even in terms of recent activities.
The delivery of a .25 Athlon was a big deal, and making the transition to .18, and having .18 scale in mhz were two more - remember that Intel was showing roadmaps indicating Athlon classic wouldn't make it past 700mhz.
But the big delivery wasn't Athlon, it was copper at Dresden - Paul, Elmer, and everyone else at Intel, including Barrett and Otellini, were absolutely convinced that AMD wouldn't be able to yield worth a darn on copper. That's why they let themselves get so far behind in competitive capacity.
Then they thought AMD wouldn't be able to produce chips with on die cache to keep up with coppermine's cache.
Then they thought AMD wouldn't be able to design and ship a dual processor platform - and AMD shipped one that substantially outperformed Intel's best.
Then they thought AMD wouldn't be able to design and ship a CPU with "smart" power saving circuitry - then AMD shipped Power Now that exposed Intel's speed step as the nearly useless system that it was (at least, by comparison).
Then they thought AMD's power consumption would hold them back - and that was a tough problem - but the engineers at AMD came through with the Palomino series.
In the same time period Intel came up with some good work such as coppermine's fast on-die cache, but they also tried to shove Rambus down the throat of the computing industry, and they are continuing to push Itanics ridiculous architecture. P4 was mediocre at introduction, and is looking better after some fixes - but no more than so-so (so far).
Before you have a stroke, remember that the very large, just released, P4 Northwood, using a lot of silicon on a very expensive brand new process, has about the same performance as the smaller Palomino, that has been shipping since June (as MP) and uses an older .18 process. |
