I've had this board for about a week now and some further impressions...
I don't care for Foxconn's implementation of their dual BIOS setup. This board has a mind of its own in that it will boot off of the backup BIOS whenever it feels like it. I had to update to the latest version of the Foxconn Liveupdate program to be able to flash the board's backup BIOS to the same version as the primary BIOS. I've got both flashed on the latest version now, much better! I don't know if this board actually uses two BIOS chips or if it places the backup BIOS on the hard drive
(I've read conflicting things) but if they do indeed use two chips I would prefer a jumper or DIP switch for selecting which BIOS to use instead of having the board automatically choose. Also, I would've preferred if the onboard power button was a CMOS reset button instead. Really that's a minor gripe, but in my case the CMOS reset jumper is a tight reach if I'm using the nearby fan header as well, a button would be easier.
I thought this board was really picky about DDR2-800 RAM as I was having trouble with getting two different sets to boot. Turns out I was just having some trouble understanding the memory dividers on this board. The dividers are listed as: 400mhz, 533mhz, 667mhz, and 800mhz. Figured out that this translates to:
400mhz = 4:3
533mhz = 1:1
667mhz = 4:5
800mhz = 2:3
I've tried a pair of 4850s in this system with DDR2-533 running the E6600 at 356x9, and the memory divider set to keep the RAM at 535. Here are some 3dm06 scores:
Two cards installed, primary card running at 8x, crossfire disabled: 3dm06 = 11578
Two cards installed, both cards running at 8x, crossfire enabled: 3dm06 = 14902
One card installed, card running at 16x: 3dm06 = 11683
Nearly 15k 3dmarks is 1k more then my overclocked 8800GT gets on my P35 running a QX6700 at 400x8 and DDR2-800 ram at stock speed and timings.
edit: Now that I have the dividers figured out I have the system running at 356x9 with some DDR2-800 and the divider at 1:1. Here are some 3dmark06 results of that with and without crossfire active:
crossfire active, both cards running at 8x: 3dm06 = 15089
crossfire disabled, primary card at 8x: 3dm06 = 11648
I saw some nice improvements with enabling crossfire, particularly so in 3dmark03 where the score went from nearly 41k to over 71k seen
here and
here.