Originally Posted by Burn
Sources always back one's arguments up, but I think I can place at least some trust in The_Manual's arguments.
I do understand the importance you place on NB straps. If the 400mhz strap on the P5B wasn't designed to run at 425, why is one able to do that on the BadAxe? I also don't think I am going to play around at 425 FSB much, I am hoping at least 4 Ghz, if not more. So around the 450-500 FSB range hopefully.
The E6600 has a 9x multiplier locked both ways, correct?
On the P5B imagine a chart. The chart has a a series of FSB and latencies associated with it. When you hit 400Mhz FSB, the NB initiates the latencies under 400Mhz on that chart.
This chart would be just like the timing table on CPUz.
So for the P5B, if you step over 400Mhz FSB, you are now in a new range of latencies. You have no control over this.
However, on the XBX2, you DO have control over it. You can change when the latencies change. Which allows you to overclock to 400Mhz + FSB and have a huge advantage in performance over the P5B.
If that isn't enough to explain it to you, take a look at this:
Here is a 1M run comparing a
XBX2 @ 450Mhz x 9 = 4050Mhz 1800Mhz FSB
P5B Dlx @ 503Mhz x 8 = 4025Mhz 2012Mhz FSB
Even with the FSB and memory advantage, the XBX2 is faster!
Again, same setup, look at the Aquamark03 score.
Here, the XBX2 is also faster than the P5B Dlx:
So as you can see, the ability to manually control the reference frequency (NB strap) allows for more performance!
Also, the board has better power regulation.
Owners of the board swear it is more stable and overall better than any ASUS board.
I may get one (going to be hard justifying 3 boards in as many months, but this is just an amazing board!)