D9's 6-6-6 1800+ no way......

[I.nfraR.ed]

EDIT: Are those #'s on air Infra?
Reguardless, pretty damm impressive bud!
I'd be more than happy with those results.

Thanks, Dave . Water on the cpu and air on board and rams.

That NB speed is outragous, I dunno what you are complaining about :p
(/rant on) My 965 did 3 Ghz NB exact through 32M, wasn't prime stable even at 1.45v. My 565 doesn't like to go over 2850, doesn't matter what volts, it doesn't scale over 1.35v.
My 1100T does 3 Ghz NB stable at 1.35v, but even booting 3.1 on water is impossible at 1.4v.
Those ram clocks are crazy for 32M stable on a 955 at 3:10 divider if I'm not mistaken.
Also, none of my CPU's do 32M above 4300 >.<
Darn GTX2 >.>, my STT sticks only do 32M up to about 1870 6-6-6-18 at 3:10 without cold. (/rant off)

This 555 is from the newer Deneb with 3 letters. It does easily 4.5GHz, even Cinebench stable on normal ambients, 4.7GHz 3DMark 01 is what I've tested on it. CPU-NB is ok, but 400Mhz behind Thuban (mine can run 3.6GHz on 2 cores). Ambient temp was ~15C, required to be stable on this CPU-NB speed. But you're right, my STT can't do that, at least on air.

Well you certainly proved the fact that 890 should be used with thuban,High clocks on 890 are because it runs slack

NB gains aren't worth that much, some but maybe 1 sec per multi up.

Yeah, it seems it's not efficient. I pulled down 2 seconds from going 1880 6-6-6 to 1940 6-7-6 and another 2 seconds with 1932 6-6-6. With my lack of skills it's the best I can do atm. Such high clock on 890FX isn't very easy, it required higher voltage than I usually use. Let's say I used "enough" volts for these timings
I would love to test with Crosshair III, but I sold mine and that particular board actually didn't clock the memory at all.
The other one I have at the moment is M4A89GTD, not sure about its efficiency, but it should be similar to C4F.