Tenchusatsu,
You haven't seen PCI-X, and it wasn't even an Intel invention. But who cares about elegance, as long as it gets the job done? AGP gets the job done (though not in the way Intel originally intended), and so will PCI-X.
Regarding PCI, I think it was still the old Intel, when the techies were still running the place. I don't know much about PCI-X, but it looks like it will be good as well.
But AGP sucks at every respect. It's premise is flawed, the execution is flawed, and just about everything related to it is bad. The premise of a main computer memory being used in high performance graphics systems for storage for the graphics card is downright comical. Connecting AGP to the Northbridge is just about the ugliest design one can think off. All of the cards that were produced in either PCI and AGP showed minimal to no performance improvement for AGP.
The only rational reason for AGP (assuming that there was some logic to it) were either the monopolistic reasons, or to create churn, force an un-natural upgrade cycle.
In my opinion, Microsoft should hike up the specs on X-Box and choose a faster processor in order to stand a real chance of taking on PS2. A 733 MHz Pentium III will be on the order of a microcontroller by the end of 2001. Instead, X-Box should really be sporting a Pentium 4. By the end of 2001 (or mid-2002), P4 should be mainstream and cheap enough.
How about X-box-2? I doubt about P4, since it will still be too power hungry, but the Tualatin versions may be coming. I don't think there is a chance Sony will be able to keep up performance wise with the x86 + nVidia + MSFT combination. These 3 companies can leverage everything good that goes to the entire Wintel architecture, that may be worth $100s of billions annually.
The only thing in Sony's favor is that most people will be still connecting these consoles to the same low quality TVs that we use today, somewhat reducing the potential advantage of X-Box.
Joe |
