Used to be able to get 4.7ghz on 1.3v(1.312 irl), now it's fucking up even at 4.5ghz.
Stress tests are fine, nothing crashes. Now things will freeze randomly, not even under stress. Watching youtube, all that shit.
Any help on what to do?
I also have the intel tuning plan, should I just kill thing and get a chip? Anyone done this before? and how are the replacement chip?
Check for info about the batch you got.
If there isnt any info, kill it and get replacement.
>14nm FinFET
>1.3V
Yeah, totally not elecromigration killing you're shit slowly.
Being stable and cool doesn't mean it won't damage itself over time.
ive run my 4790k on 1.3V @ 4.9GHz for like 3 years... no crashes and works perfectly.
It's also not built on a 14nm process.
its not on 14 nm...
Shit, thank god i didint replace my 4790k with a 6700k then...
welp, so I might as well suck a cock and get a new CPU then?
>that core voltage
You're definitely degrading your silicon. It's been noted that Skylake materials are "less" than Devil's Canyon and even that had a max safe ceiling of 1.3.
Thats a lot of voltage for the chip. It could possibly be the VRM on your mobo, but you might have done some small amount of damage to the CPU.
what would you say I should have it at?
I followed a guide on overclock.net and they said you can go up to 1.40V if you had water.
>1.4
What the fuck.
Wait a few months and get a real CPU (AMD Zen).
my motherboard is a ASUS ROG VIII Hero, been said to have the best VRM out there.
have I been memed and shilled to?
Thats for max possible OC.
It is not in anyway safe for a reliable system you plan on running every day and keeping for years.
The shills got ya.
and throw away my $200 cpu with my $200 motherboard?
haha, yeah okay man.
Sell it on eBay
fuccccccccccccccccccc,
I fucked up bad then...
What voltage would you say I should have it at?
Keeping this for about 3 years.
>1.4
Dumb fuck good luck wasting your money for new CPU+main board.
I'll be able to get $300 max for both. That's not counting shipping and packing which will be stupidly expensive.
thx dad, what would you say to keep it at then?
Probably set it back to stock.
A 20% clock uplift wasn't worth inducing a lasting hardware fault.
1.4v is a lot for 14nm, even pushing Sandy Bridge that hard wasn't exactly recommended, lower the voltage and deal with a lower frequency if you want stability...
it's totally stable under a stress test, ran 100 cycles of IBT on max (14GB).
I'm just not sure if I fucked up the cpu or something else.
>Now it's fucking up at 4.5
>Stable
what voltage should a 6700k be at? mines at 1.3v and constantly changing due to the turbo that i havent turned off yet (6700k is only a few days old)
you can, but those posters also replace their CPUs every year or so
You can literally Google "boardwell-e 1.45v" and see people running the latest 14nm bwe processors at 1.4v or higher.
Core voltage doesn't kill processors, cache and vccin voltage does.
But my guess is that your processor wasn't stable to begin with. Next time use occt.
4.7ghz is nearing the top end of skylake and requires extensive tweaking. Sorry to say this but if you're asking this question then you likely didn't tweak it right.
That image is very familiar.
>Core voltage doesn't kill processors
Also, as I said earlier here Even if it runs stable and cool on such a high voltage, doesn't mean that it will continue doing so for months.
You simply can't predict the effect electromigration will have on a chip over time with the little anecdotal evidence you have as an enduser.
I guess you could be right, but for it to drop 0.2GHz at the same voltage and still not be stable I don't buy it. I've ran it for over 3 months at 4.7 and then I had to back it down to 4.6.
Only in the last few days I've needed to turn it down to 4.5ghz.
Something is dying but I'm not sure what it is.
Try and stay at or below 1.26v for 24/7 and also disable load line calibration (LLC) if you make use of speedstep (EIST in BIOS) as the two in combination can/will cause damaging voltage spikes when a power state switch happens going into full load (the spikes are so fast they are often undetectable in software tools)
I would dial back the OC or investigate if it is some kind of memory instability (did you overclock the IMC?), in my experience core instability will throw BSODs whereas northbridge can cause programs locking up and "chugging" like you described
i rember reading that with skylake electromigration decay was getting pretty real for the first time since a long time
That's not much higher than stock voltage, you dumb shit. Skymeme uses higher voltages than Haswell across the board.
Thats not a good thing.
Intel had consistent yield issues with their 14nm process starting with Broadwell CoreM. The knee of the curve for peak perf/watt is significantly lower than stock clocks for Skylake. Pushing vcore is the only way intel could bin sufficient chips to reach their target clocks, a standard practice in an underyielding IC.
The chips simply won't last as long when running at a higher voltage.
>CPUs with built-in expiration date
Intel's crowning achievement
>Stock core voltage on 6700k can go to 1.37v
>you're shit
>you're
IBT is borderline useless for stability testing on OC'd Intel CPUs from Haswell onward. I mean, it'll catch obviously unstable shit but other than that you're better off playing some CPU heavy game for 30min than running IBT for 999999 hours and murdering your CPU with extreme temperature and voltage.
Try the OCCT stress test too and if Skylake is anything like Haswell you will want to play around with the ring bus clock/voltage too, as well as memory related settings. Try keeping the ring bus at stock and use an officially supported memory clock (not XMP).
In 2 years OCN, Anandtech, and other forums will be full of posters talking about their overclocked systems failing to post and only running at stock or lower clocks.
What should be preemptively call it?
The Skylake is falling? Kry baby lake-gate?
I'll work on the logo
So whats a proper voltage to run Skylake at assuming you overclocked a 6600k to 4.5GHz? Or rather, when does the decay really start to kick in?
The truth is nobody really knows (except Intel, and they're not telling) and it also depends on other factors such as temperature.
I mean, shit already runs at 1.25v stock. So putting at to 1.275v or 1.3v really SHOULDNT be a problem. Would be mind blowing if it was to be honest.
Yes, going by "common sense" increasing the voltage from 1.25 to 1.275 or even 1.3 shouldn't be much of an issue, if you keep the CPU decently cool during your daily workload.
You are mistaken voltage with current.
Keep the temps in check and you'll be fine. Current is like a big cock that goes back and forth and degrades processors.
>You are mistaken voltage with current.
Increasing your CPUs supply voltage will in the end result in increased current too, however.
so what ?
is it bad?
AMD's 32nm at 5GHz by default
Wait a second. Someone help me: the voltage on my CPU is supposed to be capped at 1.25V yet for some reason HWM shows that it reaches up to 1.31V from time to time. What? Is that another faulty shit like Speccy not working on half the components?
The first thing I do when I get a CPU is to see how far I can undervolt it and remain stable on stock. After you do that, THEN you start your overclocking.
No idea why people are going mental in here. I ruin a 6.6k at 4.5 and never even touched the voltage. Checking CPUID and HWMonitor reveals, that with the Core Voltage being on AUTO, it reaches as high as 1.31V from time to time on its own. And, again, thats running the DEFAULT values of my motherboard. So I kinda doubt that you did anything wrong, might just be the tech being shit.
If anything try it all over again. I like to offset just a tiny bit of voltage and then cram the clockspeed to impossible levels. Then I just lower by 100 MHz until it is stable. I'm on 4.4GHz with a voltage of what appears to be 1.2v. Linux sensors say it's on about 60C at load, so I could push it further with a little more voltage.
My 4790k used to run 4.8 NIggahurtz but after about a year and a half I started tripping the over voltage protection. Turns out it had climbed to about 1.38v at peak.
Had to drop it to 4.6 where it sits stable on 1.25v.
Skylake can take higher voltage than all other generations due to the IVR being off-chip for once. Have been running mine at 1.33 for 4.6 since when I got it in September. Runs fine. Maybe it;s the shitty Asus mobo?
tfw i run my cpu at 4.5ghz @ 1.17v totally stable
4790k a best cpu
but dat shit runs like a 2GHz Celeron