why doesn't Intel spec the next AGP revision to run on a faster bus - not a quad-pumped 66mhz bus or whatnot, but just raise it up to 100mhz? I mean, it's pretty much a given that AGP4x gives basically no new benefits on current systems, while raising the AGP bus does. La Intella could design the new bus to have some kind of different connector that's backwards compatible with AGP2x (like AGP4x is now), but have the extra connector tell the motherboard that the card is spec'd to run @100mhz. That way, current 66mhz bus cards could run on the new connector, while allowing the new cards to run at 100mhz.

Sure, designers would have to deal with signaling issues and all that, but wouldn't that kind of design change give raw, usuable performance gains, as opposed to the theoretical bandwidth given by stuff like AGP4x? I can't imagine it being any harder to implement such a design than it was to implement the PC100 and PC133 buses on motherboards.

Now, I haven't read the info on AGP8x in a while, and Intel may have something like I said in the works, and I may be talking out of my *** (again), but, hey, that's one more post toward my post count

------------------
The Idiot Brigade Homepage
The Sharky Counterstrike Team meets here, for now

Milhouse: We could put it on the internet.
Bart: No! We have to reach people whose opinions actually matter!

[This message has been edited by tu2thepoo (edited August 05, 2001).]