K8's don't have a FSB so to say, so increasing to 250MHZ would have done very little. You don't have to increase the FSB to increase memory bandwidth on the K8, you can do that by lowering the frequency divisor.

Anyway HT always works in multiples of 200MHZ, so something like 250MHZ doesn't work.

Anyway those Anandtech benches are quite old, not to mention you don't have a baseline for compairson, we don't know what the clock rates was for the system running at DDR2-800, they don't provide that information becuase they were still under NDA at the time.

Increasing the FSB is necessary on Conroe since the processor could still use more bandwidth with the limit being 8.5gb/s with 1066FSB, even utilizing Dual Channel DDR2-800, with AMD since they use IMC they get the full 12.8GB/s with Dual Channel DDR2-800, though both processor won't reach their theoretical rates.

And currently has no processor on 333FSB. That is something we will be looking for in the future at the earliest Q2 2007, with the intorduction of Bearlake and Weybridge chipsets.

Yes you are right.

I thought they where testing at stock speeds since they only put the memory timmings there.

Intel should remove the extra cache and amd should remove the IMC HT that way we know which chip is better clock for clock. :)

The Intel Conroe have EIST which drops the frequency down to 1.6GHZ, when the processor is idling, since they are limited to x6 multipliers.

Thanks for that info. Do you have any idea what the power draw might be at 1.6GHz? Will it be comparable to the power draw of CnQ at 1GHz?

I smell cheese here.
Both are water cooled so therfore its the processor? Cheese
A AMD3200 running at 2400MHZ and with the quote "overcloacking is for idiots." Yep more cheese
I ditch your power supply first.

More flashy = More hits?