MSI Z270 Motherboards Automate Core i7-7700K Overclock to 5.20 GHz

[Outback Bronze]

Aren't all modern CPU's running at higher voltages than old ones? I've seen several Skylakes running at way higher voltages than the old ones...

Well I thought/heard it was the other way around. Voltage should go down as the manufacturing process goes down. The lower the nm the more fragile it becomes apparently. But it seems not in this case. Surely MSI wouldn't risk this setting in the bios if they didn't think it was plausible. They are putting themselves at a hell of a lot risk and the processors along with it. Like I say, maybe they know something we don't or I don't....

 

[proxuser]

Same story every year. I was able to test 3 6700k and not a single one made 4.8 stable.

 

[EarthDog]

If you plan to keep the CPU only for less than 1 year, than 1.5V is fine.

want to bet it will last longer than a year?

Now, I wouldn't run it like that, but intel.intel . spec is 1.52V max. I've run chips at intel max for years...now, this isn't penryn/45nm, but it's going to last longer than a year with the appropriate cooling and board.

Really?

High end air cooler > AIO liquid CPU cooler

lol, not anything decent with 2x120mm radiator...

 

[Outback Bronze]

want to bet it will last longer than a year?

Now, I wouldn't run it like that, but intel.intel . spec is 1.52V max. I've run chips at intel max for years...now, this isn't penryn/45nm, but it's going to last longer than a year with the appropriate cooling and board.

lol, not anything decent with 2x120mm radiator...

Had a mate run a i7 920 @1.6v for a couple of years. Didn't break. It had custom water cooling tho.

 

[Solidstate89]

Really?

High end air cooler > AIO liquid CPU cooler

Nope. Not unless you erroneously believe the only AIO setups out there are 120mm ones and that's it. My 280mm one is better than any high-end air cooler.

 

[LTUGamer]

Nope. Not unless you erroneously believe the only AIO setups out there are 120mm ones and that's it. My 280mm one is better than any high-end air cooler.

Really? As hardware reviewer three years ago I have made roundup that includes 6x LCS
(Cooler Master Seidon 240M, Corsair H100i, Enermax Liqtech 240, NZXT Kraken X60, Silverstone Tundra TD02, Thermaltake Water 3.0 Extreme) shows similar performance to D14 and nobody could beat D14 in noise/ perfomance ratio. Newer tested coolers such as H105, H110i GTX or Silverstone TD02-E won't changed situation

 

[EarthDog]

We aren't talking noise to performance ratio... just performance.

 

[Supercrit]

Aren't all modern CPU's running at higher voltages than old ones? I've seen several Skylakes running at way higher voltages than the old ones...

Smaller process means bigger susceptibility to electromigration, due to less material to resist the degradation of circuits by the currents.

I had a friend's Nehalem i7 that could reach 4ghz (maintaining 24/7) at a higher voltage but as time goes by, the CPU could no longer reach that speed without crashing, very possible due to electromigration damage.

 

[LTUGamer]

We aren't talking noise to performance ratio... just performance.

Performance is nothing. When fans spins on 2000 rpm system noise is uncomfortable. The best cooler is that one which offers best performance at 1000~1500RPM

 

[mcraygsx]

Performance is nothing. When fans spins on 2000 rpm system noise is uncomfortable. The best cooler is that one which offers best performance at 1000~1500RPM

Actually Noctua Nh-D15 is perfect for noise discipline.

I noticed these KABY Lake (even some skylake) need a high voltage to overclock at and beyond 5GHz. My own 7700k need 1.38 Voltage @ 5Ghz. But this 1.5+ Vcore is intense. Wish these tiny chips had larger die size like Extreme Edition ones for sake of thermal contact.

 

[EarthDog]

Performance is nothing. When fans spins on 2000 rpm system noise is uncomfortable. The best cooler is that one which offers best performance at 1000~1500RPM

Performance is what we are talking about though. You are welcome to change the subject as you see fit to make your point... but it isn't what the discussion was about.

 

[Breit]

A lot of people say 1.5 will kill your CPU faster, but do anyone have some evidence that it wont still last 3+ years?

Can't say nothing about Kaby Lake, but my i7-990x died after ~2 years@1.4v first with demanding slightly more voltage every now and then to be stable with the overclock I had no problems with for the 2 years. I recon it was 4.6GHz back then. not exactly sure though.
And then all of a sudden it wouldn't boot at all with that chip, no matter the voltage and settings (even stock).

The same is now starting to creep up with my current Sandy Bridge E 6-core, which I have clocked to 4.8GHz on all cores for ~5 years now. It started about a year ago, where the chip suddenly demanded a slightly higher vCore of 1.4v (was 1.38v for ~4 years prime-stable and all). I'm now at 1.42v using the exact same overclock.

This Sandy Bridge E system probably lasted longer because I implemented an offset overclock opposed to the 990x, where the CPU varies its voltage from ~0.85v while sitting idle on the Windows desktop up to 1.42v when doing actual work. This also keeps energy bills, heat and noise in check.

All that is with a hugely oversized watercooling setup (4*480 rads, 2 pumps, everything watercooled), so no temperature issues whatsoever.

Being that said, I would say sure, hard overclocking does have an impact on component life. The question is, if the system should last you longer than the reduced lifetime of your CPU due to overclocking or if you have any problems with switching CPUs if that is happening to you sooner.
In my case I'm fine with it and the performance gain justifies it for me personally.

Back to the topic: I'd be interested in the power draw of this CPU at those voltages. My system draws ~300W more compared to idle only when I start number crunchers such as prime (no GPU, just CPU, VRM and RAM). I only could imagine how crazy this would be at 1.5v... oO
Needs to be a very good AIO though.

 

[Recon-UK]

Just leaving this here.

 

[Enterprise24]

1.52V is not that too scary on Skylake/Kabylake. I have been use i5-6500 @ 5Ghz 1.52V for nearly 9 months without any sign of degradation (I use my PC everyday 6-12 hours/day). Skylake non K OC cannot enter power saving state. So it run at 5Ghz 1.52V all time even just web browsing. So the i7-7700K that can enter idle state should have no problem if you can COOL it.

 

[GhostRyder]

Cool that it will auto do it, bad that its so high on the voltage front.

I don't believe those Intel processors should be pushed above 1.4 (really, I don't even like beyond 1.3 much for 24/7).

 

[EarthDog]

You can likely tweak it down Ghost. As with most all boards, when using AUTO settings for voltage, it tends to go high in favor of stability. It will of course depend on the CPU sample and how good it is, but.. I'm sure you get the point.

For example MSI boards, Z270, set SA and IO voltage when using XMP to a sky high 1.45 and 1.4V. To run 3866 I only need 1.25 and 1.2V respectively.

1.3/1.4V is nothing... I would happily rock 1.4V all day long until the thing died years later... seriously.

 

[R-T-B]

You sure like throwing away money.

Upgrading CPU every 2 years is pointless.

No, I like playing with hardware. Who are you to judge if that's throwing away money or not? I enjoy it, therefore it's worth it to me.

Why do they disable C-states, I thought that C-state was for idle power saving....odd people

Probably because when extreme OCing they affect stability, and power saving is the last thing on their minds?

 

[GhostRyder]

You can likely tweak it down Ghost. As with most all boards, when using AUTO settings for voltage, it tends to go high in favor of stability. It will of course depend on the CPU sample and how good it is, but.. I'm sure you get the point.

For example MSI boards, Z270, set SA and IO voltage when using XMP to a sky high 1.45 and 1.4V. To run 3866 I only need 1.25 and 1.2V respectively.

1.3/1.4V is nothing... I would happily rock 1.4V all day long until the thing died years later... seriously.

Nah, I get that but I just meant for someone just doing "Auto" is a little high especially if the person does not realize that's alot of voltage. Its more I think it allows to much on "Auto" than I think it should.

As far as 1.4, its just a personal preference. I don't see 1.35 being horrible, just a little bit more for my taste but I know its not going to end its life in the next year from that.

 

[eidairaman1]

Final Run tRC 31 1866
http://img.techpowerup.org/170108/capture003-20170108.jpg

tRC 30 1866
http://img.techpowerup.org/170109/capture005-20170108.jpg

This a Turbo Core setting or across all Cores?

PS check my voltage and max temps lol

 

[R-T-B]

As far as 1.4, its just a personal preference. I don't see 1.35 being horrible, just a little bit more for my taste but I know its not going to end its life in the next year from that.

Personally, I see anything as up to 1.5v as safe, since Intel litterally lets the chips request up to 1.52v automatically via VID. They literally rate them for that. I don't know where people get these cautionary numbers, but provided you can keep the chip cool, they make no sense.

 

[damric]

Haters gonna hate

 

[Jism]

Any respected overclocker should always do manual oc'ing and testing. Leaving this up a predefined auto-system is asking for troubles.

You know, oc'ing once started with litres of coffee in the middle of the night testing, and now this auto program comes around just booting up to 1.5V in an Intel CPU like it's nothing, lol.

That CPU will proberly have serious stability issues with 1 to 2 years.

 

[TheinsanegamerN]

Cool that it will auto do it, bad that its so high on the voltage front.

I don't believe those Intel processors should be pushed above 1.4 (really, I don't even like beyond 1.3 much for 24/7).

1.3V? heck, my I5 runs at 1.015@4.2, and gets a bit uncomfortable warm at 1.1v, which is stock for 3.8.

Seems to me these 5GHz @1.5V suffer pretty badly from diminishing returns.

I also remember a friend having a phenom X6 on a MSI board try to auto OC, and it overvolted both the CPU and the chipset. No surprise, it was dead a year and a half later.

 

[ADHDGAMING]

A lot less impressive considering the fact that someone already broke the 5Ghz mark 2 yrs ago, without even needing much OC to get there.

 

[EdInk]

Out of curiosity has anyone here run a CPU at higher than average 24/7 and seen it go bang. If you have experienced this yourself(not hearsay) please let me know. I'm getting sick of seeing comments about above average voltage bad.

What are the tolerances of the transistors in these CPU's?