I'm glad you elaborate a bit more on your setup, it's certainly a bonus in a review. But can you say that they (AT) are wrong?
You investigate the difference between the FRAPS and timedemo times, but what about the difference between your timedemo and theirs?
If you are trying to prove your...
F5A has nothing new according to a poster at this page:
http://www.forumdeluxx.de/forum/showthread.php?t=356358&page=37
Apparently F4A didn't fix the PS2 problems to Gigabyte's satisfaction.
Could still be cpu volts, mine needed more to reach more common overclocks.
Though, I haven't done the vdroop mod yet. It's next on my list ;) (thx btw fishton for the info)
IS: You know that using onboard video will limit your overclocking, right?
Push your FSB up gradually, and try increasing the voltage of each of your components (in turn) when you find problems. Or go down from your current unstable OC until you get it stable.
It's most likely the NB/MCH isn't...
Yeah, I thought it was strange that you needed both to, but then gigabyte boards do have their quirks.
I'm running F4d now. All I can notice different is that memory speed isn't reported at bootup, and CPU-Z reports more infos (memory SPD I don't think was working for me before)
Spent a bit more time investigating. The higher mem multi for fsb800 cpus seems a bit dodgy. I have tested at fsb329 (highest stable POST, now I'm working out voltages for cpu & NB) with the 3.33 multi and mem at 2v and memory passes memtest fine. That's ddr1090, when ddr1080 wouldn't work at...
2.7 is as far as I've tested so far, haven't had time to work with it more, actively cool my ram etc. But it's 27hours dual prime stable at that, and plays my games so nicely...
I'll probably push further when I get my next day off.
I posted a request to Gigabyte for 1:1 on FSB800 processors. No reply yet. *fingers crossed*
Not that I can really complain, my e2160 has gotten to 2.7 on stock volts, and my Corsair XMS5400-c4 has gotten to 500mhz on 1.9v.