Why BX is faster than KX133 and 815

[Helevitia]

I wanted to mention this since it seems obvious to me but maybe I'm missing something? Isn't the BX chipset faster because the AGP bus is already overclocked to 89Mhz? Both the KX133 and the 815 have /2 dividers which leaves the AGP bus right where it should be at 66MHz, right?

Dave

[DarkSyd]

You are correct about the AGP Dave. As for why the BX is faster, I don't know what the exact technical reasons are, but simply that BX still seems to be architecturally superior, particularly in the design of the memory subsytem.
I'm sure someone else here can give you the nitty gritty details.
It sounds like you're doing a similar upgrade to me. I went with a FCPGA Coppermine 700 on an Asus CUBX. It should all be here tomorrow, but my new fan and heatsink will be a few more days. I'll let you know how it works out.

James

[paulcs]

I certainly think the overclocked AGP slot helps. I've heard some mixed reports about the i815e. Some reviewers have had it outperforming the BX chipset by a small margin, while others have old reliable on top.

I just think it's too soon to tell. I suspect people are using different board revisions, drivers, BIOS settings, etc.

There are some ominous theories about that the i815/815e is artificially underachieving so as not to be a threat to the i820.

I suspect the VIA Apollo Pro 133A is just plain flakey, and we'll see better performance from VIA's upcoming revision.

Paul
paulcs@flashcom.net

[Nuno]

815 is slower than BX because Intel decided so. It would be embarassing is their low-end chipset managed to out-perform the 820 RDRAM chipset.

815 has a hub architecture, which in theory has much more bandwidth available than the classic northbridge/southbridge. It *should* perform better.



[This message has been edited by Nuno (edited 22 June 2000).]