Will going above 1.4v on a delided i7 7700k and Kraken x62 be too dangerous?

[NiBy]

i lost the silicone lottery and and i want to hit the 5ghz on my 7700k

i think i should be able to achieve it at 1.4 to 1.435

im just a bit iffy about going above that 1.4 mark.

my cpu is delidded with thermal grizzly conductonaut.

i uses thermal grizzly kryonaut in between the ihs and aio

here are my temps after running realbench for about 30mins to let the liquid heat up

https://imgur.com/a/1CgY43x

 

[Karadjgne]

I too lost the silicon lottery, on my i5-3570k. It'll sit very nicely at 4.3GHz at 1.08v. 4.4GHz or above is unstable at any voltage higher, upto and including 1.5v at any LLC level, ram setting, c-states or eco setting or with all board currents set to 255 max.

Will 1.4v damage that cpu? Yep, seriously shorten it's usable lifespan. Will it be stable at 1.4v or higher? Nobody can say, only the cpu. And it's not the temps that do the damage, it's the voltage inside the cores burning out nodes, creating electrical bleed etc.

 

[cookiemania66]

Your temps look okay but that run was with 1.2V and 4.5Ghz. You can go beyond 1.4V to an absolut max of 1.45 for kinda Daily use.
The question is more like: will that 100mhz be worth the XX more power consumption and probably reduced life span?

Going above 1.4 an your chip is going to bo above 80 i guess.

 

[larrycumming]

not worth it... just run it at the sweet spot overclock.

 

[geofelt]

The danger comes from excessive voltage.
1.4 would be my limit.

Delidding helps to control temperatures at a given voltage, but you are more likely to run out of safe voltage before you run into thermal limits.
The processor will slow down or shut off if it detects a dangerous temperature; that is around 100c.

You always hear about wonderful overclocks, but those with dogs are silent.
I think your chip is average.

As of 6/9/17
What percent can get an overclock at a somewhat sane 1.4v Vcore.

I7-7700K
4.9 83%
5.0 62%
5.1 29%
5.2 6%

 

[dederedmi5plus]

Did you use der8auer delid die mate 2? on Z270 Maximus Apex or expensive OC mobos? did you follow video tutorial https://youtu.be/DX24ocSJ4AI by der8auer? did you plan to use https://www.ekwb.com/shop/ek-sf3d-inflection-point-evo ?

The more VRM phases a motherboard pack, the more likely it'll maintain 1.3V+ to 1.4V+ stable operation, your success 5GHz OC for me only convince me that all K series 7th and 8th gen intel CPUs are 5GHz OC-able with good cooling and delid.

What I haven't figured out is "how long a 1.4v on 5GHz 7700K endure", 30 min is too short, maybe you can present us with more useful work such as POV Ray or Cinebench?

 

[Karadjgne]

30mins on an Aio is about right to acclimate the coolant. Aircoolers need less than 10mins and most of that is just for case heating purposes.

And no. While more likely, silicon lottery is still rampant in Intel cpu's and there will still be a good sized percentage that will simply not be stable at 5.0GHz no matter what voltage or cooling. It's due to impurities in the silicon itself and will not only change from batch to batch but even placement of the cutting, whether it's towards/at the edge or center. Center cuts show a considerably lower rate of impurities as the manufacturing process tends to push most of the heavier impurities like nickel, copper, etc towards the edge.

Power phase count helps, but only to a certain extent. Higher phase counts tend to lessen the impact of vdroop, but won't eliminate it altogether, you'll just end up with a larger amount of waves per period on the DC line. The largest benefit to higher phases comes when paired with a better psu that already has a somewhat cleaner, more stable output. Basically the better Gold and Platinum psus. For most Bronze users, higher phase counts beyond 12 or so really don't do much at all.

 

[dederedmi5plus]

30mins on an Aio is about right to acclimate the coolant. Aircoolers need less than 10mins and most of that is just for case heating purposes.

True!

And no. While more likely, silicon lottery is still rampant in Intel cpu's and there will still be a good sized percentage that will simply not be stable at 5.0GHz no matter what voltage or cooling. It's due to impurities in the silicon itself and will not only change from batch to batch but even placement of the cutting, whether it's towards/at the edge or center. Center cuts show a considerably lower rate of impurities as the manufacturing process tends to push most of the heavier impurities like nickel, copper, etc towards the edge.

Care to elaborate that?

• ■ I alway thought core count is the limiting factor of clock rate ( more core = less MHz, less core = more MHz), if 4 cores can hit 5GHz then why 6 windows parked 10 cores i9 7900X can't hit with the same cooler? or
  ■ is intel use different TIM for each? why 8086K such differ from 8700K, 8700 and 8600K?

Power phase count helps, but only to a certain extent. Higher phase counts tend to lessen the impact of vdroop, but won't eliminate it altogether, you'll just end up with a larger amount of waves per period on the DC line. The largest benefit to higher phases comes when paired with a better psu that already has a somewhat cleaner, more stable output. Basically the better Gold and Platinum psus. For most Bronze users, higher phase counts beyond 12 or so really don't do much at all.

PSU supply the CPU with 12V EPS 8 pin > the mobo's PWM switch that 12V to user's selected value, AFAIK on consumer PSU no +12V rail was reported drop below 25% by online reviewers, would the +12V voltage drop will still be over the 1.4V vcore ripple threshold,

 

[larrycumming]

this has more to do with design and production efficiency. for example, haswell-e processors are shipped either 6-core or 8-core, but are of the same physical design layout.

6-core variants are produced in the same physical layout but with 2 of the 8 cores disabled. thus, you'll see the 6-core chips run thermally better than the 8-cores because of the two "cold" inactive core on the chip.

this isn't just an intel thing, AMD is no exception.

30mins on an Aio is about right to acclimate the coolant. Aircoolers need less than 10mins and most of that is just for case heating purposes.

True!

And no. While more likely, silicon lottery is still rampant in Intel cpu's and there will still be a good sized percentage that will simply not be stable at 5.0GHz no matter what voltage or cooling. It's due to impurities in the silicon itself and will not only change from batch to batch but even placement of the cutting, whether it's towards/at the edge or center. Center cuts show a considerably lower rate of impurities as the manufacturing process tends to push most of the heavier impurities like nickel, copper, etc towards the edge.

Care to elaborate that?

• ■ I alway thought core count is the limiting factor of clock rate ( more core = less MHz, less core = more MHz), if 4 cores can hit 5GHz then why 6 windows parked 10 cores i9 7900X can't hit with the same cooler? or
  ■ is intel use different TIM for each? why 8086K such differ from 8700K, 8700 and 8600K?

Power phase count helps, but only to a certain extent. Higher phase counts tend to lessen the impact of vdroop, but won't eliminate it altogether, you'll just end up with a larger amount of waves per period on the DC line. The largest benefit to higher phases comes when paired with a better psu that already has a somewhat cleaner, more stable output. Basically the better Gold and Platinum psus. For most Bronze users, higher phase counts beyond 12 or so really don't do much at all.

PSU supply the CPU with 12V EPS 8 pin > the mobo's PWM switch that 12V to user's selected value, AFAIK on consumer PSU no +12V rail was reported drop below 25% by online reviewers, would the +12V voltage drop will still be over the 1.4V vcore ripple threshold,

 

[larrycumming]

double post

 

[Karadjgne]

4core i5-3570k and 4core i5-3340 both have 4 cores. Got nothing to do with MHz. Has everything to do with power used inside on the cores. Both use @75w in their 1155 pin setup upto 3.8GHz. The i9-7900x is a 2666 pin, 140w monster with 10 cores that can hit 4.3GHz. Totally different power usage, and it's consequent heat output.

Cpus are regulated by physical limitations, like the cooling available. With the sheer amount of heat generated by 10 core use (that 140w is not a measure of heat, but a measure of the power used under a nominal load of a Intel specified series of apps) the cores run slower on multi-core use to prevent the cpu from reaching critical temps. Not uncommon for massive core count Xeons to run @1.8GHz or so just because of the heat generated that cannot be dissipated by conventional means.

I7-8086k is an anniversary edition i7-8700k that's been binned by Intel to remove the silicon lottery as a factor when getting the cpu to 5GHz