UnOfficial Asus Rampage II Extreme Thread

[modUlar]

Oh man, I had my chip on 1.6v vcore briefly like 15-20s, will this damage my chip? It was on 1.3 vcore, but then I changed the absolute VID to relative VID offset or whatever to see what would happen. And after POST, I went right back into BIOS and saw the vcore was like 1.6v or something, so I freaked out and immediately changed it back to absolute VID and 1.3v. Do you guys know how long it will take a i7 to be damaged by high vcore? I'm very worried right now.

lol relax man your chip is fine

[KILLER_K]

I noticed that for my kingston hyper x 2,000 that i need 1.62+ qpi to boot at 4.0ghrz and over. If anything else lower it will not boot in windows it will simply restart. It must be nice for all you getting those high bclks as mine wont do nothing hardly over 4.3ghrz. Trying to figure out what is the best way to go with it now. Something somewhere is walling me bad. It is now more annoying then anything i guess.

So up for suggestions at this point. I would like to get a higher bclk on this chip but perhaps it isn't going to happen. Plus since i placed this kingston back in the settings are all weird now to get it to even boot up.

[akeedthe]

So im trying to OC the first time...i got my sig rig a week back...
idle temps are 48C-50C & prime95 temps are 64C...
ambient is 34C...
NB & SB temps are 60C & 58C respectively
are these temps reasonable considering my ambient even with a megahalem?

i dont want a 4ghz OC ( it would be nice tho) if my temps allow me to do so...

My 1600 ram wasnt detected as 1600 but as 1066 so i had to make timings of 8-8-8-24 & dram bus voltage to 1.65 (stock settings all)

can i pls be directed on OCing my system...
thnx

[KILLER_K]

Should be easy to get 4 ghrz with a d0 and that huge megahalems. You can try these settings if you like and if you feel you ram is holding you back raise the multi to 21 and back off the bclk some.

Here try these settings.

Extreme Tweaker

1) AI Overclocker Tuner [manual]
2) CPU Ratio Setting [20x]
3) CPU Turbo Power Limiter [Disable]
4) BCLK Frequency [200]
5) Dram Frequency [1600]
* keep this close to your ram specs as you can
6) CPU Voltage [1.40]

Save and exit and see if it goes into windows. If it does you need to test for stability. If not give it a little more Cpu voltage and the max i would go on a c0 would be 1.45 volts. As i'm not sure if you have a co or d0 stepping. Use these apps to get your info.

CPU-Z
http://www.cpuid.com/download/cpuz/cpuz_152_setup.exe
* Use this to get your cpu speed , stepping , motherboard, ram and video card info.

Core-Temp
http://www.alcpu.com/CoreTemp/
*not sure if you have 32bit or 64bit os.
* use this to get your temps on your cpu under load running linx.

Linx 0.6.0 beta
http://www.youwatched.com/datajay/linx(0.62).zip
* use this to test the stability of the overclock. If it passes the 20x loops then you are fine. If it blue screens or restarts up the voltage on the cpu volts in the bios.

Also when you look at cpu you will see your cpu not showing the full 4.0 ghrz clocks. That is because some settings in the bios slows down the cpu wen not under a load. Some like this features others don't. so if you don't like it do these steps to make it run 4.0ghrz all the time. But note your temps will rise some after this.

And if the temps get too high, you can ......
a) Buy a better cpu cooler
b) Disable HT in the cpu options in the bios, same place as i'm fixing to show you under this.

When you disable HT it stops using to threads for each core, making it perform like a regular 4 core cpu and not a logical 8 core cpu, 4cores x 2 threads = 8 logical. This will knock 15° - 20° off your full load temps and some off idle temps, refer to core-temp for that information.

CPU Settings
1) C1E Support [disable]
2) CPU ™ Function [disable]
3) INTEL SpeedStep Tech [disable]

These 3 settings will stop the cpu from slowing down when not in use, i personally like it as it save power and reduces heat. But to each his/her own.

There are a few more factors that will help you along the way but this should get you started. Rest you will have to learn by trail and error. As each piece of hardware overclocks different. So test things out and do some benchmarks, write down the results until you find the sweet spot.

Some benchmarks are.
Vantage, if you run vista.
3DMark 06
Cinebench
Sisoft Sandra
Everest

These are a few of the good ones. Maybe this will help you along and you probably will not need 1.4 volts since you have a d0 so you can try lower if you like.

[akeedthe]

so ill start off with 1.4v on the cpu eh? i have a D0 also...
wat abt QPi voltage???

just another question...when CMOS is cleared using the button at the back...does the bios version get reverted as well?

[akeedthe]

so i used settings given by KILLER K...booted at 4ghz an ran n passed linx 20x...
but temps are YIKES!!!!
check out max temps!!! theres a 100 there!!!!

[unclewebb]

See that LOG word in the Thermal Status area of RealTemp? That's a sign that you triggered thermal throttling. When that turns to HOT, that means thermal throttling is presently in progress. When it first starts to happen it will just momentarily throttle your multiplier lower towards 12.0 until the temps get under control. This can happen hundreds of times a second. It's interesting to see the core beside it got up to 99C but didn't trigger any throttling. You know what the max of your chip is now.

You might be better off running 21 X 190. This gets you similar total MHz but most users can do this more reliably and with less core voltage and heat.

Save LinX testing for the winter time. 34C ambient is a bit too warm for this kind of testing but nice to see your overclock is stable.

If you want to see what your multiplier is up to during thermal throttling then try running i7 Turbo. It's included in the latest RealTemp download.

http://www.fileden.com/files/2008/3/...alTempBeta.zip

[KILLER_K]

lower the cpu volts as you have a d0 and it shouldn't require that much volts. Just keep lowering it and testing it with linX and see if it is stable as you should be able to drop the volts a good bit on it. As i already see you have Ht disabled, if the temps keep getting hot even after you drop the cpu volts down. perhaps you got a bad seat on the heatsink.

Try using core temp also it seems to be a little better then real temp for me.

[unclewebb]

Try using core temp also it seems to be a little better then real temp for me.

I realize that I'm a little biased here but there should be no real difference on Core i7 CPUs. Both programs correctly read TJMax from the processor and they both read the exact same sensors. The only difference is that Core Temp ignores APIC ID information which means on some motherboards, if your bios does not set your APIC ID in order, Core Temp will not associate the reported temperature with the correct core. Send me a PM if you have a specific issue with RealTemp or something that it is telling you that you disagree with.

[akeedthe]

k...so i got 4ghz on 1.25v...load temps at 80C....is that good?
or should i try to go lower on cpu voltage & see?
also...is there anyway to save the BIOS file on the RIIe?...coz if i dont post when lowering voltage i can just use my prev settings

[KILLER_K]

When you are done with the settings go into the bios.
* asus o.c. profile
* add your cmos profile
highlight name it
* save to
highlight the profile you like to save it to

You should see profile change to what you named it to. Save and exit. You can save it to a flash drive etc so if you flash the bios to a newer one, you can import those saved profiles back into the bios.

[Mr.Pink]

that is not a wise idea to import saved OC profile to a new bios

most ppl have reported that when importing a saved profile the oc is no longer stable
just my

[akeedthe]

4ghz...prime p5 small fft 5 hours stable...

http://valid.canardpc.com/show_oc.php?id=658848

[scaramonga]

that is not a wise idea to import saved OC profile to a new bios

most ppl have reported that when importing a saved profile the oc is no longer stable
just my

Which makes it a kinda useless feature then.

[exe163]

A question about Linx. Is it normal that after each loop the CPU downclock itself and then goes back up to 4.0? I am trying to get a stable 4.0ghz with C0 920, but each time it downclock my temp drop significantly down to near idle. Only for a second or so, will it effect the accuracy of the test?

Which makes it a kinda useless feature then.

Well if your BIOS crashes, it saves a lot of time re-inputing all the data in.

[zoqiet]

A question about Linx. Is it normal that after each loop the CPU downclock itself and then goes back up to 4.0? I am trying to get a stable 4.0ghz with C0 920, but each time it downclock my temp drop significantly down to near idle. Only for a second or so, will it effect the accuracy of the test?

Well if your BIOS crashes, it saves a lot of time re-inputing all the data in.

If your CPU down clock its because of your power saving settings.
Buuut, the linX test works like this; Every time it posts results the cpu load goes down to 0% then it restarts the calculations and go up to 100% again. So it will drop in temperature near to idle and its intended.

OC Profiles are excellent, I use them a lot when I swap around playing with memory or cpu benching.

[exe163]

If your CPU down clock its because of your power saving settings.
Buuut, the linX test works like this; Every time it posts results the cpu load goes down to 0% then it restarts the calculations and go up to 100% again. So it will drop in temperature near to idle and its intended.

OC Profiles are excellent, I use them a lot when I swap around playing with memory or cpu benching.

So Linx is basically 5 minute prime with 20 stops? How could this be a stress test when it only goes for 20 seconds and then reset itself? Most of the errors happens 6+ hours into prime test, unless I am missing something I can hardly imagine calling a system stable after only running the loops once. ~6min

[zoqiet]

So Linx is basically 5 minute prime with 20 stops? How could this be a stress test when it only goes for 20 seconds and then reset itself? Most of the errors happens 6+ hours into prime test, unless I am missing something I can hardly imagine calling a system stable after only running the loops once. ~6min

It's pretty much the same, at 4ghz + one calculation takes about 200seconds. (LinX, 6gb ram)
In prime95 it just keeps on feeding calculation after calculation, its still just a bunch of of them (its not one single big calculation), all queued up after each other, prime uses different sizes of calculations. Well you can specify it. Small, Large, Blend etc etc.

Difference with linX is that you just choose one size to calculate with.
And in between it stops down a bit, then go back on. It's intended and I would presume its because of vdroop. For checking if your system is stable with your volt going up/down during min and full load.
For a proper stability test I'd recommend using more than just 1 application though.

[kup]

And also, LinX and Intel Burn Test both use Linpack which is INTEL's own inhouse CPU testing application so it can't be all that bad.

[dejanh]

So Linx is basically 5 minute prime with 20 stops? How could this be a stress test when it only goes for 20 seconds and then reset itself? Most of the errors happens 6+ hours into prime test, unless I am missing something I can hardly imagine calling a system stable after only running the loops once. ~6min

That's wrong. A single pass in LinX (Linpack) with 5GB+ DRAM will take roughly 3 to 4 minutes at ~4.5GHz on an i7 with HT on. Overall, the test takes 1h+ to run 20 passes at those clocks and in the meantime it will push the temperatures of your chip on every pass well beyond those that Prime95 will generate. That's what makes LinX very good for testing stability quickly vs. running Prime95 for a freaking day.

Anyway, this is a R2E thread so let's get back on topic.

[Leeghoofd]

Well I tested the latest LinX 6.02 and compared it with Prime Custom 20K, temp differences are 5°C on my WC setup. So wayyy beyond is no longer the case like it was with S775, 10°C and more...tested on the same WC kit here (just other mounts for the CPU block). This ofcourse taking into consideration that some air coolers loose grip at a certain temp and the watercooling allows to go that tiny bit further... but the question remains ; does a test that generates more heat is more conclusive in determining stability ? I can tell you it's not !

And this is the best quote of the day : "For a proper stability test I'd recommend using more than just 1 application though."

[Sizmik]

Hi everyone, I've been dying to post but it took a while to get my account activated, but I'm all good now.

what I wanted to know is that when you guys update your BIOS, did you save a profile first, default the settings, update the bios, then load your profile back on ? Mine looks to have worked fine but is that the best way of doing it ? Also, in my BIOS I states that I am actually running at 4GHZ but cpuid says I'm at 4.2GHz.

[Bodkin]

Hi, just got one of these boards, much better than my gigabyte which died. However coretemp reads about 10 degrees higher than the LCD poster temp. Which should i trust?

[kuup]

My D0 validated at 238 Bclk. I think it may have a little more in it if I could get it a little cooler.
I really don’t know the maximum Bclk that anyone has achieved on a REII though.

i can get mine @228 bclk rock solid n benchable but gettin a instant bsod on 229.

CPUZ
spi
max BCLK

cant find any stability with clocks over 4.4Ghz HT on r off

[THXTEX]

Hi everyone,

Just wanted to say hi to you all. Built my new rig a month ago but had some stability issues. I looks like one of my memorysticks is bad. I got errors with both Goldmemory and Memtest. Will have to wait for replacement before I can start OC. Are Goldmemory and Memtest still the best to test RAM?