• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

5ghz

Guest2

Guest2

Guest2

Soldato
Joined
6 May 2009
Posts
20,235
I have a new i7 4790k and it clocks pretty well, so far at 4.7ghz, with 1.220 core voltage.

Max temps in IBT is 82c this is using a noctua nh-u14s.

Should I try for 5ghz, possibly with 1.29? Does anyone know the max temps for these chips

Thanks
 
Are you stable at 4.7 with that voltage? That is really low.

Have you only ran intel burn test? How many passes?

It's always worth running more than one stress test, as I've always found sometimes you can be "stable" in one stress test and bsod in others.

A fully stable cpu should pass a mulitutde of stress tests. If you are fully stable with that voltage then you've got a good cpu.

Stress testing is always different for people though, I know some that overclock and provided they can run their games without crashing they're "stable".
 
Last edited:
Looks pretty ok, but once you up the voltage, temps will get much higher using ibt. Id suggest using Asus real bench to test with. Much more realistic than ibt/p95.
 
Mine is stable at 4.7GHz @ 1.214v vid according to CPUz, 1.224 actual vcore (8 hours realbench)

I havnt finished testing above this, id probably want to delid if I did, but 4.8GHz is NOT stable at 1.25v, and 4.9GHz is not stable at 1.3v. In fact they not even close to stable really, they crash within 5 mins of stress testing.
 
Last edited:
new boy said:
Mine is stable at 4.7GHz @ 1.214v vid according to CPUz, 1.224 actual vcore (8 hours realbench)

I havnt finished testing above this, but 4.8GHz is NOT stable at 1.25v, and 4.9GHz is not stable at 1.3v. In fact they not even close to stable really, they crash within 5 mins of stress testing.
Click to expand...

That's the thing, sometimes you can just hit a brick wall.

Cooling can help obviously, when on air temps are the brick wall, but even with a good watercooling loop you just hit the brick wall of the silicon lottery.

Also you get to a point where a couple of 100mhz won't give any real world gains anyway, but it does feel good having a cpu sitting close to 5ghz :D

You probably already know this but it's not only vcore that needs adjusting when overclocking at the higher frequencies, so perhaps you are able to get stable at 4.8ghz.
 
Yeah, I've been playing with input voltage too. Dropped the ram to 1600MHz for testing so shouldnt need sa or digital/analog IO voltage adjustments, but left them at what's needed for 2400MHz at 4.7GHz anyway.

I'm sorely tempted to delid. I'd defiantly get 4.8GHz at a 24/7 level of voltage and temps then, 4.9 would probably be ok too, I'm sure it'll do it at under 1.35v at least.

But as you say, its obviously not worth it from any kind of logic perspective...

EDIT

On topic, there doesn't seam to be a consensus on temps from what I can tell. Possibly because of the wildly different temperatures different stress test programs give?

Personally, I felt uncomfortable when I saw 85c in realbench and stopped pushing after that. During that heatwave we had a few weeks ago, I was hitting 83c in realbench @ 4.7GHz, now more like 80c. I'm happy there.
 
Last edited:
Thanks.
Is it better to leave C3 on to lower voltage when not loaded and also leave EIST so the frequence isnt at 4.7 (or whatever speed) when idle?

I have always disabled them before but surely its better for the chip if they run cool when not needed

idle temps at 800mhz are around 30c
 
I've left all power saving stuff on on mine + using offset voltage. Doesn't seam to have sold me short on the overclock as far as I can tell.

I think in general people have found it not to effect haswell overclocking at all, so why not right?

I leave my PC on 24/7, dont really no why but I do. So 99% of the time its at 800MHz @ 0.8v
 
Guest2 said:
Thanks.
Is it better to leave C3 on to lower voltage when not loaded and also leave EIST so the frequence isnt at 4.7 (or whatever speed) when idle?

I have always disabled them before but surely its better for the chip if they run cool when not needed

idle temps at 800mhz are around 30c
Click to expand...

I personally always disabled c3.

I'm not particularly bothered about power savings, and I found with high frequency overclocks having it enabled could cause instability. Plus I just like having "max power" all the time :D

When overclocking as you know it shortens the life of components, but to be honest by the time it has any effect you'll have likely moved onto a new rig anyway.

I know many have said that leaving any of the c states enabled doesn't have an effect on overclocks the way it used to years ago, but from personal experience I've always had less issues disabling it. That and running less sticks of ram.


new boy said:
On topic, there doesn't seam to be a consensus on temps from what I can tell. Possibly because of the wildly different temperatures different stress test programs give?

Personally, I felt uncomfortable when I saw 85c in realbench and stopped pushing after that. During that heatwave we had a few weeks ago, I was hitting 83c in realbench @ 4.7GHz, now more like 80c. I'm happy there.
Click to expand...

Stress tests are real worst case scenarios, in normal usage temps would never really get that high. Add to that the fact that it's really hard to kill a cpu, your rig would shut down if it got close to a dangerous temperature anyway.

Intel cpus are nowhere near as sensitive to heat as amd cpus for example. In saying all that though I wouldn't ever run a really high overclock on air, it's all about watercooling for me, where you can stress test to the max and not go over 60c with a real good custom loop.

I used to use a triple rad running push pull to cool just a cpu lol. I just honestly can't be bothered with the maintenance nowadays, I'd always check the loop, res etc to make sure it wasn't leaking, I'd spend more time doing that, and then became obsessed with lowering temps, lapping cpus, using different thermal compound, different fan configurations etc than actually using it.
 
Last edited:
Running ram at 1600mhz is not a downgrade I think the difference between running 1600mhz and higher is minimal.
Running ram at over 1600mhz can effect your overclock peak and volts required.
Check out the haswell guide http://www.overclock.net/t/1411077/haswell-overclocking-guide-with-statistics full of good info such as clocking the uncore isn't required.

2 games I find fall over pretty fast if unstable are arma 3 and left4dead2 20 vs 20 hentai server great for stability test and many more of course.
And on haswell I read it can dangerously overvolt if you stress test on adaptive voltage on some boards.
Also read up on your ram as 1 timing looser could be the difference between stable and unstable at chosen voltage and clock.
After finding clock use all power saving states on balanced power scheme in windows and adaptive voltage.
 
Last edited:
Thats weird, coretemp and realtemp are now showing max of 100c when running IBT!

Core voltage is still at 1.200v. Could the temp sensors be wrong or something else pushing up the temps? This is at 4.6ghz (46 x 100 on manual multi, no memory overclock)

I could understand if it was crashing due to not enought voltage or if it was on auto or set to 1.4v or something but its not :confused:
 
Real Bench is great I agree.

Intel xtu also as its what Intel recommends.

Using your pc for what u actually need it stable for is the best stability test.
 
The fan on my Noctua NH-U14S doesnt sound like it ramps up at all even when running tests, is this normal or is it just a pretty quiet fan even at max rpm?
 
Any motherboard software running to monitor fan speed? Personally I use x2 2350 rpm corsair sp PWM fans on my k2. But use the turbo profile in bios fan control. I can certainly hear them when they ramp up. Iirc, on Asus boards once the CPU temp hits 60c fans run at max.
 
Hmmm, something is wrong somewhere. I hae just removed the overclock and gone back to stock settings. Running Intel burn test at maximum increases temps to 100c and fan cpu fan speed goes to max - 1480RPM

It suddenly jumps from 66c > 100c when running IBT

Voltage reported in CPUZ is 1.236. Do you think the max temps are incorrect or could it be to do with something else such as XMP or another setting?

Motherboard is Gigabyte Z97X-Gaming 7 with the latest F7 bios

edit - its also a new HSF with new Arctic MX4 applied

Thanks
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom