• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Upgrading 775

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Looking good!

Just remember that even if the voltage is higher at idle it's not anywhere near as hard on the CPU as a lower voltage under full load. Under full load is when the current goes way up and electromigration occurs faster.

Don't be afraid to experiment with your settings. That's the only way to really figure it out. Unless someone has an identical mobo or very similar it's going to be hard to give very specific advice. It's up to you to try things out, take notes, and see how different settings affect the end result. ;)

You are on the right track! Keep it up!
 
:D That's good to hear that you guys approve and that Jason has a larger form factor of my board :p
You'll see how BIOS roughly looks like, key difference is that the BIOS differes slightly between the two - being that your current version uefi version is 2.30 and mine is 1.30.. But it shouldn't be that much of an issue.

On the side:
My PC is way too noisy to leave it on for the night and I don't tend to do that but I will run more tests once I get back home, seemed stable yesterday (eventually).
I shall try and push for more.
 
There're other voltage settings you can fiddle with, but only if your board let's you change them.
What voltage options do you have? Only offset?
If so, you're screwed pretty much.
 
Oh i have a few other voltage settings... Actually there is a whole tab under the O/C Tweak Tab (a UEFI tab just for o/c :D HOw thoughtful of Asrock :)) devoted just to voltages. There are like 8 or 9 settings there but I have no idea what they all do 0.0 I will need to take a screenshot to give you an idea.

Besides you don't think that a brand new socket 150$ motherboard with Extreme written on it would lack any necessary O/C settings? :p

However, I just went up to +0.075v and i'm on my 20th test, no errors. at +0.070v one core failed almost instantly. Damn, these things have fine tolerances

EDIT:
1 self test passed. 15 minutes, not much but It appears to be pretty stable. I think i will keep it at this speed for an hour or so and play some games, see if anything comes up. If not then I will restart and push for more.

BTW, While in Prime95 voltages kept at 1.344-1.356v
While opening Firefox I hit 1.392v?! Then idles around 1.072v
 
Last edited:
Okay I have BOOT'ed at x47 with +0.080v and whenever i do anything, even as much as opening Firefox or going to a different web page, voltage indicated spikes to 1.4v. I wonder which setting makes it go that high since Prime 95 tests up till now usually maxed at nearly 0.050v lower than the spike voltage.

Thing is it likes to sit at that 1.4v when not being used to the maximum of its capacity.

EDIT:
When I opened Prime95 then i could do anything when all 4 tests are running and vcore keeps to 1.352. when one fails it rises slightly to 1.358 and jumps to 1.376-1.384 with another one down the drain. Odd?

Here are some UEFI Screenies:

121020112401.jpg
121020112452.jpg
121020112504.jpg
121020112511.jpg

121020112519.jpg
121020112529.jpg
121020112542.jpg
121020112553.jpg


EDIT 2:
+0.085v x47 1 Self test 512k passed.
Looks Like i found a stable 4.7ghz O/C, lets try some more.

BTW do you guys know what fans would further increase the performance of my 212 EVO? I know that 55-65 max temp in Prime95 for a 4.7ghz core is not bad, but I want it cooler still. Noise is not a concern, but price is. I am trying to find stock EVO fans to go on the back but neither Amazon nor Ebay are helping me with that. USA CM website has one but I dread to think what the postage will be.
So I could get a full replacement for 15$ ish. OR find another matched fan within that price range? Any ideas? Anyone know where I can grab a stock one for my EVO from?

EDIT 3:
Im still looking for an answer as to why my Vcore jumps when the CPU is not being 100% used and when under full load, falls by 0.03v? Especially now that this means my Vcore goes above 1.4v and idles at 1.1v. But I don't let it Idle a lot so most of the time its around 1.4v mark... And still unstable at +0.105v x48

EDIT 4:
Vcore spikes up to 1.440v at times, with +0.110v offset.
Bios shows voltage of 1.230v
Prime95 load is now 1.384-1.392
And also Prime95 just crashed....

EDIT 5:
I have achieved a stable 4.8ghz overclock at +0.120, I believe. Prime has ran a first self test with no errors and is now running a new set of iterations - strange thing happens, the processor throttles back to 3.3ghz while running those, staying at a comfortable 1.288v.
When I do something to force the core to go into turbo while stress testing, it goes up to 1.4v 4.8ghz
 
Last edited:
Okay I need some help now.

I was stable at +0.120v offset yesterday. Today, it doesn't even pass the second test on Prime95. It crashed yesterday as I was about to switch it off with code 124 i think. So today I booted with +0.125v and it hanged during prime95. So I increased it again and it hanged even quicker. So I lowered it back to +0.120v and now it doesn't pass prime95 anymore.

What's annoying or even distressing is that my vcore is at .120v under no load but spikes up to 1.440+ and 100% load requires over 1.4v @ 4.8ghz. Apparently they should only need 1.39v at this clock.

What should I do now? Should I push for more? Should I throttle back? Leave it?

EDIT:
Tried various settings and concluded that:

1) Load line calibration should be left on 100%, otherwise spikes rise to 1.47v or MORE....
2) C1E should be left Enabled otherwise I am on constant 1.4v+
3) VDroop only exists under 100% load situations (odd?)
4) CPUID shows voltages around 0.15v DIFFERENT than BIOS. Different but in the wrong way, I don't like having 1.256v in BIOS and 1.4v in CPUID...
5) For overclocking any more than 4.7ghz I think I will switch to increasing the Turbo Voltage rather than Offset since offset adds voltage to the core AT ALL TIMES and Turbo Voltage only adds more power when called for.
6) I can idle around 1.080v (1.6ghz) which isn't a surprise but I cannot explain how 4.7ghz (clocked down since 4.8 seemed ''scary'') pushes the vcore up to 1.4v and spikes to 1.440 (if it goes any higher I might not have a chip to o/c in a moment...)
 
Last edited:
Ok, I'm checking out those bios screens now...not familiar with Asrock bioses myself.
Load line calibration is making your vcore overshoot when under load. Reduce it a bit and you should see.a tiny droop under load, which is the way it should be.
Code x124 indicates that your ram is unstable, or that you need more dram voltage, vtt, or vccsa.
Do a thing, test 4.8ghz with your current settings but leave ram at stock settings, you'll never know what causes the instability like that man. You HAVE to do things one at a time ;)

EDIT: dram voltage is still very low, I already told you not all ram modules behave the same, and not all cpu memory controllers neither. You can't just copy settings from a review site and expect everything to work...so take your time and first isolate the cpu oc, to find a stable setting. Then you can fiddle with ram. Or the other way around...
 
I suppose but I took that review as a rough guide as to what these sticks can do.
Much in the same way that everyone says 2500k boots up at 4.5ghz, give a bit more voltage to the ram and it will work at these speeds as well.
And in the review their timings were tighter than mine. But okay.

Trouble with LLC is that when I DISABLED it, the spikes actually increased. Under medium load I was hitting 1.47v..... consistently.
And I only have 100%, 50% and 0% options apart from auto. 100% being the safest. But I have no idea what else could possibly cause those spikes.

Right now im on 4.7ghz and +0.090v which was the last stable overclock since 4.8 became unpredictable.

On the side:
The more reading I do about O/C the more confused I am. I know roughly what each setting does but I still don;t know what to do with it and when to do it, apart from the basics obviously.
I also found out that CPUID core voltage is actually VTT voltage.
I found quite a few other people that have similar issues with voltages being different. BIOS Vcore is fluctuating around 1.25v when cpuid says its 1.4v+ Which one is correct? Should I be worried that Im 0.03v below the maximum safe voltage or have I got a lot more headroom?

EDIT:
HW Monitor says the same thing so core voltage indicated is the actual core voltage...

EDIT 2:
Tried some 4.8ghz again, went as far as 0.100 offset and +0.055v additional turbo voltage. Still workers kept failing almost instantly in Prime. Pushed the DRAM up to 1.475v, that stopped code 124 BSOD's for now.

In the end I clocked back to 4.7 and got rid of PLL, Additional Turbo and left the offset at 0.100v. Workers keep failing.
 
Last edited:
Those ram sticks need more voltage to run like that...Forget the review man.
I'd downclock them ASAP to stock settings and volts to perfect the cpu oc itself.
You NEED to do them one at a time, otherwise your brain will hurt :p

Do this:

1-Get rid of llc and use smaller offset values.
2-Try to get a cpu-z vcore under load under 1.45v and polish that 4.7ghz oc
3-If prime95 custom test passes, reduce vcore till it bsods. Work it from there...

Once you get the lowest possible vcore value in cpu-z with no dead workers nor bsods, you can proceed to overclock your ram.
 
What about PLL, VTT etc? Core Current Limiters?

BTW I just had to clear CMOS, due to the fact that I lowered offset to -0.030 and tried to compensate for it with turbo voltage. That got me stuck unable to enter UEFI menu or POST. Annoying thing is I can't work out whats wrong and I read that recommended PLL is around 1.750v?


EDIT:
I went back to x45 and im trying a different approach. First off, C1E, EIST, C-States, Visualisation etc. all that lot is OFF.
offset +0.030v
LLC is OFF.
Memory is back to stock.

Results?
So far no errors.
But voltage after boot is 1.392v at all times - since speed stepping is off thats not worrying, what IS worrying is the sheer amount of voltage needed for 4.5ghz without powersaving.
While running Lehmer 11000 iterations, voltage drops from 1.440+ to 1.288-1.296. Spikes go as high as 1.456v

Now what could possibly be causing the voltage to be THIS high? I can see it dropping, maybe its because of VDroop, but then it rises by over 0.1v just like that and as far as i can work out, prime95 stresses the cpu to its maximum all the same... where does the fluctuation come from now?...

EDIT 2:
Hmm I hit 76 degrees on core 3 under these conditions. Repeatedly.... I should point out that although lehmer 800000 iterations caused a down clock to 3.3ghz and only need 1.296v, in moments when starting a new set of iterations, core goes back to 45 and 1.456v, jumping straight away to 75 degrees on core 3 and up to 72 degrees on other 2 cores, core 1 goes up to 65. So far.

I don't get it. Voltages went ridiculous as soon as i switched off powersaving and LLC but even with an offset as small as 30mV the core is close to cooking itself 0.0

EDIT 3:

C1E and EIST enabled back, lowered offset to +0.025v.
Load voltage is 1.432v.... Is VDroop non existent? I appear not to experience that. More to the point, at NO point the CPUID voltage was lower than BIOS voltage - indicated by hardware monitor.
Core likes to throttle back to 3.3 on prime tests for no apparent reason.

EDIT 4:
No LLC still, upped to x47 at offset +0.025v. Voltage keeps at 1.448v, goes as far as 1.472v.

When under 1.456v+, temps soar by as much as 20 degrees almost instantly, before throttling back for another run of iterations.
But once that happens, I could even watch hi-definition videos on youtube AND be on the forums, this seems to barely affect the temperature and it still throttles back of its own accord.

EDIT 5:
Ivan, look at this: http://www.overclock.net/t/1261592/asrock-z77-how-do-i-use-load-line-calibration

My board doesn't use levels but percentages. But its asrock so its possible they turned everything upside down 0.0
 
Last edited:
Prime95 doesn't stress the cpu the same in all tests...You should base your offset values on the highest spike you noticed to be safe.

But I have no idea what to do with that information still :( Im trying to lower the offset. So far 0.020v seems to work but my Vcore under load is either ~1.288v or ~1.448v and thats not good.

Tried going for 48 but got a 124 BSOD.

Could you look at the screen i posted with all the voltage options in the OC Tweak menu and walk me through? I just don't know what to do with anything at the moment... Gonna try lowering offset still.

EDIT:

+0.010v, didnt even flinch. Hmm.
 
Last edited:
I did, it's easy. Your motherboard only supports offset overclocking and that calls for turbo enabled.
So, according to your screenshots here's what you should try:

-Disable additional turbo voltage, thermal throttling, power saving mode, load line calibration (What choices do you have under CPU Voltage offset?)
-Enable Speedstep and Turbo
-Enable spread spectrum and use 100.1 bclk
-Set the max value for Long/short duration power/current limits and long duration maintained
Try a 4.5ghz oc first, prime it and see how that goes...
Monitor max vcore under load in cpu-z.
 
I can't even touch Turbo option, its enabled by default and thats that...

Hmm I enabled all C-States and pretty much everything apart from LLC, power saving and y offset is +0.035v at the moment - that Is I was running this when I read your post. x47 for now, voltages still go mental when under load.

Under offset options I can only see values, negative and positive.
Under CPU Voltage option itself there is only Offset and Auto.

I primed the hell out of 4.5 already at various settings and seems that it can take pretty much anything. I still dont get where spikes (Ever-increasing in both voltage and frequency...!) come from. I hit 1.480v and keep teetering on the edge of 1.448-1.46 but now with ever more high voltage fluctuations....

HAd to increase VTT to stop BSOD code 124.
 
Back