***Asus Rampage III Extreme Owners Thread***

[donmarkoni]

On another note there must be a clockgen that controls the QPI PPL voltage. Does anyone know where on the board it is? Perhaps a variable resistor might be put in place to increase voltage like on video cards for example. I am not saying it would work but just an idea...

+1 It would be good to know, if we can't get it in BIOS.

That is the thing Don, it is the lowest i can go with vtt, anything lower and i get bsod 124 (uncore = 4092) and raising RAM i need even more vtt.. my chips memcontroler sucks xmas balls >_<.. but im in a early state of testing this bios.. so, now i know R3E CAN do it.. its all in the settings.

Well, it's a pretty high UnCore clock (4096MHz), so it needs much VTT. IMO, it's too much for daily use. I would search for maximum UnCore clock while keeping VTT around 1.4-1.425V (or maybe up to 1.45V).

[Fire4Wire]

...and in my system it broke the S3 Suspend to RAM wake up.

As it seems you have a preferential link with Asus BIOS development, what's about the user requested QPI PLL voltage settable and separated from CPU PLL?

nothing @all @the moment

if this option comes in further bios


my opinion NO

[donmarkoni]

nothing @all @the moment

if this option comes in further bios

my opinion NO

And what about this:

On another note there must be a clockgen that controls the QPI PPL voltage. Does anyone know where on the board it is? Perhaps a variable resistor might be put in place to increase voltage like on video cards for example. I am not saying it would work but just an idea...

Is there any way to know (from the inside sources) anything about measuring points for QPI PLL?

[tistou77]

I had a negative experience with 0095.
Pushing an overclock a little over stability edge, caused the impossibility to reboot, even unplugging AC from mains. When you pushed the ON button on the PC, it started for about one second and after it shot down by itself.
Solved removing the cmos battery for a while and clearing cmos with the rear button.

Back to 1102 for now.

you do not test the 0011 bios?

to compare it with 0095 bios

[cky2k6]

How is this board with high baseclocks and 3d? My old r2e was pretty terrible, but I hear conflicting info about this board. Does it have the same 217 ceiling the r2e had?

[kurumi_Japan]

How is this board with high baseclocks and 3d? My old r2e was pretty terrible, but I hear conflicting info about this board. Does it have the same 217 ceiling the r2e had?

Good REX2 has not these low BCLK wall.
I tesed 5 REX2, bad one faces BCLK wall under 210MHz, but good one over 235Mhz.
Mainly it depend on IOH quality, not mobo design.
But I think that REX3 average BCLK limit improve than REX2.
3D performance good same as REX2.

Anyway, I think UD7's average BCLK limit more better than REX3.
I bought 3 UD7(rev2) & these mobo BCLK limit are over 235MHz.
But UD7's memory performance is not good as REX3.
(3D performance is good as REX3.)
So you need consider what is more important for your OCing.

Only BCLK & 3D performance, I recommende UD7(rev2).
But you need memory & 3D perfprmance, I recommende REX3 or E760.
(Sorry, I had not test E770, yet.
so I do not know E770 is good mobo or not.)

I hope you will get good mobo.

[Brama]

Only BCLK & 3D performance, I recommende UD7(rev2).

The downside of UD7 rev2 is the number of phases on VRM that is lower than rev1.

A real headache to chose the right one.

[kurumi_Japan]

The downside of UD7 rev2 is the number of phases on VRM that is lower than rev1.

A real headache to chose the right one.

I hear UD7 rev2 power module are re-designed & it has enough power.
So you do not need worried about phase number.

I tested UD7 rev2 under LN2, it's very stable & can pass over 6.0GHz many benchmarks.
I think the stability is almost same as REX3.
But a little bit slower memory speed than REX3.
So it's not good for Pifast & Super PI.
Especialy SuperPI 32M is very slow, almost 10 sec behind than REX3.
wPrime is no problem.

If you have chance, try UD7 rev2.

[Rasmus66]

Im really starting to hate R3E boards...

My second R3E just died after i good weekend of benching

There was no condensation on the board and it died anyway.

Hopefully the new Gigabyte UD7 board will do the trick

[zoson]

+1 It would be good to know, if we can't get it in BIOS.

I second this. I've got a ton of pots sitting around. I don't see why ASUS won't release information like this.

I don't see why they won't release information on how a BIOS setting changes other settings in the background either.

Well, it's a pretty high UnCore clock (4096MHz), so it needs much VTT. IMO, it's too much for daily use. I would search for maximum UnCore clock while keeping VTT around 1.4-1.425V (or maybe up to 1.45V).

I'm running 12GB 2000MHz 9-9-9-25-1T with 3x4GB mushkin sticks. For stable operation at 143 BCLK I need 1.4875v vQPI.
However, for 200BCLK I do not have to increase vQPI at all for 2000MHz settings. 210BCLK my chip can't run uclk so high, but only needs 1.5125 vQPI with 1683MHz 7-8-7-25-1T.

[OC Nub]

Im really starting to hate R3E boards...

Join the club

[flesheatinvirus]

It seems to be a rather large club. I have some killer bench runs on this board at the expense of a full time shrink to baby sit its moody times.

[LardArse]

QPI PLL = CPU PLL on R3E, same source
ie, you adjust 1.45v CPU PLL = 1.45v qpi pll as well, plane is tied together so there is no way to seperate.
It did not have an effect on oc for qpi pll last i tried ; what are you trying to accomplish with qpi pll?

[OC Nub]

@LardArse, Maybe you can work your magic with this board like you did the Classified. Or at least make sure the Rex4 doesnt suffer from some of the issues many had with this board. Mainly high bclk for 3D benching.

[donmarkoni]

QPI PLL = CPU PLL on R3E, same source
ie, you adjust 1.45v CPU PLL = 1.45v qpi pll as well, plane is tied together so there is no way to seperate.
It did not have an effect on oc for qpi pll last i tried ; what are you trying to accomplish with qpi pll?

So, you say that QPI PLL (which default is 1.1V) is set at 1.8V, when CPU PLL is left at default (1.8V). It would be insane IMO.

[[200]]

QPI PLL = CPU PLL on R3E, same source
ie, you adjust 1.45v CPU PLL = 1.45v qpi pll as well, plane is tied together so there is no way to seperate.
It did not have an effect on oc for qpi pll last i tried ; what are you trying to accomplish with qpi pll?

I was hoping you would post

It's my understanding that it assists in higher bclk stability.

[LardArse]

So, you say that QPI PLL (which default is 1.1V) is set at 1.8V, when CPU PLL is left at default (1.8V). It would be insane IMO.

A terminology confusion

VTT = QPI/DRAM voltage = default 1.1v

QPI PLL and CPU PLL default is 1.80v.

[LardArse]

I was hoping you would post

It's my understanding that it assists in higher bclk stability.

No, that is not what assists bclk stability

[[200]]

Interesting, thanks again for replying.

Would you know any possible reasons why a few owners of this board struggle to reach the same type of QPI link speed as compared to other manufacturer boards like the Classified or UD range in the Gigabyte line up? I've noticed a few people mention that the IOH chips might not be as binned as well as they could be but would like to think that it might be for other reasons such as voltage settings, cooling or even the CPU itself...

Very interested in getting your opinion & thanks once again.

[donmarkoni]

A terminology confusion

VTT = QPI/DRAM voltage = default 1.1v

QPI PLL and CPU PLL default is 1.80v.

I didn't misunderstood you, AFAIK QPI PLL default is 1.1V.

[LardArse]

I didn't misunderstood you, AFAIK QPI PLL default is 1.1V.

on which brand's naming convention?

[xdave]

on which brand's naming convention?

ok Lardarse look at this guys postin the 4.5ghz stable club.... you can clearly see that the default for QPI PLL is 1.1V


Uploaded with ImageShack.us

[LardArse]

ok Lardarse look at this guys postin the 4.5ghz stable club.... you can clearly see that the default for QPI PLL is 1.1V


Uploaded with ImageShack.us

ah ok, i dont know if its just a naming clash, i meant what is called 'qpi pll' on evga MBs default is 1.80v

[OC Nub]

On the EVGA board IOH PLL is set 1.80 default, not QPI PLL. That is default at 1.1 I believe.

[LardArse]

hi default spec'ed by intel is 1.80v for at least what is called by the same name on ASUS. So if want to lower , have to lower cpu pll.