1.2GHz Athlon, 800MHz Duron
Shugashack has compiled a nice little list of sites who have already benchmarked the new chips announced from AMD today. Unlike Motorola, AMD has announced the release, not the theoretical feasibility in a few years time. I know it hurts, but they now even match the speed (in MHz) of the Xtrem Mac, which turned out to be Vaporware™, as we expected. At least there has been no word from the guys for a long time.
Interestingly enough the new Duron 800 is almost as fast as the Athlon 800, and that for only 100$.
Chips, trouble and theories
MacCentral has a sad little story talking about how the PowerPC is falling behind etc. I have discussed this very topic often enough on here, so I just point you to their article and, more importantly to the readers opinions on the bottom of the page. However, let me state one or two more things: Switching to a different chip, be it Intel or AMD, is not actually a solution for the problem. Yes they are faster than Macs now in 90% of the cases or so ( I know there are still a lot of people defending the PowerPC, and mainly the G4 with Altivec, but I stopped doing this. Just look at the games performance and you'll see what I mean. Send flames to the address over there in the frame.). The sad thruth is that these chips too will one day run into a final barrier where they will no longer be able to squeeze out even more MHz (which is by the way still the normal peoples measure of speed, Apple it's time to notice that!) and will run into the same troubles as everybody else.
I recently ran upon a strange theory I developed myself. It says that every 4th generation of a chip platform will run into trouble with either: upping the MHz, or production yields, or both. Let me point out a few things for that: 68000, 68020, 68030, 68040 and bang, Apple had to switch to a new chip.
Pentium (including MMX), Pentium Pro, Pentium 2, Pentium 3 and bang, Intel is not able to successfully push the chip beyond 1.13GHz. The development of the P4 already has been delayed and plagued with problems too. Let's go on: 601 (G1), 603 and 604 (G2), 750, 740 etc (G3) and 7400 and co (G4) and bang AIM hasn't been able to get the chip beyond 500MHz in reasonable quantities/yields for more than a year.
Now don't think AMD has never had problems, actually they had the same thing: K5, K6, K6-2, K6-3 and bang, had to cancel that line of chips because they couldn't raise the speeds. Athlon, or the K7, including its small brother, the Duron, actually belong to a new platform (namely it's a new design). That may be why they are currently the only company who doesn't face a lot of trouble getting things going (well they had a few hickups too).
I call this theory the "Blaser theory", much like Moore's Law. Whether it's a law or not remains to be seen.
Systemlogic.net has an article explaining the principles of caching, the different kind of cache and everything else you ever wanted to know about cache. Tech stuff.