Discussion Comet Lake Intel's new Core i9-10900K runs at over 90C, even with liquid cooling TweakTown

[amrnuke]

Long story short:

If you plan to spend $750-1000 on a CPU, mobo, and cooler -- you have a lot of great options. Pick the one that suits your needs and has the characteristics you care about.

 

[lobz]

My apologies if this video has already been posted, I don't have the time to check through 22 pages!

Summary for those for can't be bothered watching the whole video:

(Kraken Z63 280mm AIO, CB R20 loops for stress testing)
10900K stock - 46C
10900K 4.9GHz - 61C
10900K 5.2GHz - 91C !!
('auto' voltage)
10900K 5.2GHz - 65C
('tuned' voltage)

Moral of the story? Don't let the mobo 'auto tune' your voltage when you overclock. I can attest to this personally.

The 10900K doesn't run 'hot', but it still consumes a ton of power nonetheless, especially overclocked.

I also agree with jayz2cents comment about it taking Intel getting spanked by AMD before they even implemented these improvements to the die thickness / IHS. Too little, too late, but hey, you still have the worlds fastest 'gaming' CPU

Please let me fix your post for you. Moral of the story: the golden sample tested by Jay, didn't run 'hot'. Also the quotation marks are missing when talking about Cinebench loops as 'stress testing'.

 

[epsilon84]

I honestly believe any 10900K running close to 100 degrees under load is being 'auto OCed' by mobos that completely ignore power limits and apply too much voltage at the same time.

There is no way a stock 10900K should hit those temps if adhering to the power limits. 125W is a piece of cake for any half decent AIO, heck even a decent HSF should handle that with ease.

 

[DrMrLordX]

Temps don't matter as much as heat flux. I don't really understand why people miss this fact.

 

[Magic Carpet]

@epsilon84

Yeah, temps aren't that bad unless you run Prime95 AVX2 all the time or MCE turned on or similar.

 

[Topweasel]

Please let me fix your post for you. Moral of the story: the golden sample tested by Jay, didn't run 'hot'. Also the quotation marks are missing when talking about Cinebench loops as 'stress testing'.

That's not all its missing either. An actual torture test sure would have been much better. But he is running a brand new $150-$170 cooler, in an open platform (which regardless of what he says has a measured impact on cooling). But it ignored the actual issue with the chip in regards to power usage.

I think and the reference to hot and cold which are aspect of several things. He proved the one thing that Intel hid till launch. That they shaved a bunch off the top of the CPU to help it channel heat to the cooler quicker. Which explains why it doesn't immediately fry itself when it quickly turbo's up. Till the cooler struggles a 10900k is going to probably be cooler then a Ryzen at similar power usages. Problem is the power usage isn't equal and doing some tiered cooler testing would have been prudent.

 

[coercitiv]

Till the cooler struggles a 10900k is going to probably be cooler then a Ryzen at similar power usages. Problem is the power usage isn't equal and doing some tiered cooler testing would have been prudent.

As an additional explanation for some who read your comment and might struggle to understand the difference, shaved die leads to improved heat transfer to the heatsink, which in turn leads to effectively increasing the TDP of the cooler at ISO temperatures.

The same physical cooler, running with the same fan speed, will become stronger while running on top of 10600K instead of 8700K. This is absolutely a positive thing, but also yet another reason to remember that power usage and thermals have a complicated relationship.

 

[mopardude87]

Temps don't matter as much as heat flux. I don't really understand why people miss this fact.

I am just sitting here hoping Intel picks up on the idea of graphene lol. IBM has done it already, Intel has the resources. Then we could laugh at this. Oh hopefully in the next decade we will see it.

 

[DrMrLordX]

I am just sitting here hoping Intel picks up on the idea of graphene lol. IBM has done it already, Intel has the resources. Then we could laugh at this. Oh hopefully in the next decade we will see it.

Graphene isn't going to magically make 250W of heat easy to dissipate. It all needs to end at a radiator/fin stack somewhere. Materials with higher thermal conductivity help you get it there faster/with lower delta T.

 

[mopardude87]

Graphene isn't going to magically make 250W of heat easy to dissipate. It all needs to end at a radiator/fin stack somewhere. Materials with higher thermal conductivity help you get it there faster/with lower delta T.

General statement more then anything,but yeah i get the idea on you're saying. Yeah was on the mind, one of my friends come over and we always chat about it. Thinking more about efficiency at the processor end if it was made of another material. I saw one of the charts showing a 10900k at 99cel. Was a eye opener.

 

[maddie]

Graphene isn't going to magically make 250W of heat easy to dissipate. It all needs to end at a radiator/fin stack somewhere. Materials with higher thermal conductivity help you get it there faster/with lower delta T.

Don't worry, they expecting room temp superconductors soon. I banish you, delta T.

 

[DrMrLordX]

Don't worry, they expecting room temp superconductors soon. I banish you, delta T.

Yup, any day now . . .

 

[aigomorla]

Graphene isn't going to magically make 250W of heat easy to dissipate. It all needs to end at a radiator/fin stack somewhere. Materials with higher thermal conductivity help you get it there faster/with lower delta T.

There is a lot of issues with Graphene.

OCZ has played with it. And Tony said these to me on XS after they were done with it.

Graphene is hydrophobic....
Graphene is extremely brittle
Graphene does not like to be bonded with metal, because again its very brittle.

So u cant even use water cooling for it with a radiator.
Throw heatpipes and fins on it without difficulty because it will shatter.
And if u drop it or add too much pressure on a small area, it will crack and shatter.

OCZ did a lot of experimentation with it.
The conclusion was its too much of a headache and profit loss to do anything with it.


A better solution is something like IC Cooling Pad.

https://www.amazon.com/Innovation-Cooling-Graphite-Thermal-Pad/dp/B07CKVW18G

Its not as great as a thermal paste, but it has amazing X axis transfer properties.
That means that little foot print die heat source can be spread out to a larger foot print, and then moved more efficiently.

Only problem again, the Y axis transfer is not as great as direct contact.
But again because of its X axis spread, it does get almost near results as basic thermal paste.

Don't worry, they expecting room temp superconductors soon. I banish you, delta T.

LMAO... Even then you still cant cheat thermodynamics...
Then were gonna have people say,

"i accidentally put my laptop on my lap, and its killing my battery"

 

[Zucker2k]

OCZ did a lot of experimentation with it.
The conclusion was its too much of a headache and profit loss to do anything with it.

Wasn't it meant more as a replacement for silicon in chip production?

 

[aigomorla]

Wasn't it meant more as a replacement for silicon in chip production?

Yeah, but thermal properties of Graphene are also insane being i think 3000-5000W/mK.
Compare that to copper at 385W/mK, you can see why OCZ wanted to desperately make a heatsink out of it.

You cant just slap a graphene die over a copper IHS.
You would have a thermal runaway or a meltdown so to speak, basically the copper could not keep up with the graphene putting out, and would eventually melt the PCB, unless thermal throttle was in place, think back to AMD Duron cpus.

And Graphene is not anywhere near a room temp super conductor.
I recall Intel tried to work with graphene a while back ago, and they came to the conclusion it was just too difficult and expensive to make CPU dies out of them.

This is intel saying its too costly to make, back when they would dump millions on stupid experimental stuff just for prospecting case.

 

[mopardude87]

I recall Intel tried to work with graphene a while back ago, and they came to the conclusion it was just too difficult and expensive to make CPU dies out of them.

This is intel saying its too costly to make, back when they would dump millions on stupid experimental stuff just for prospecting case.

Oh this brings a tear to my eye, i think its the future. What a absolute waste. So much potential waiting to be untapped i got a feeling we will come back to it.

Edit:sorry for the off topic, well i guess its sorta on topic given elements always power technology.

 

[Zucker2k]

Wasn't it meant more as a replacement for silicon in chip production?

So, this is what Intel is exploring?

Intel to use Nanowire/Nanoribbon Transistors in Volume ‘in Five Years’

www.anandtech.com

 

[peonyu]

It's a economically sound CPU if you live in the arctic at least, it can double as a indoor heater.

 

[ondma]

It's a economically sound CPU if you live in the arctic at least, it can double as a indoor heater.

Getting tired of this joke. You do realize that even a room heater uses >1000 watts on high, right?

 

[therealmongo]

emotional investment in a piece of silicon is bad for your mental health

 

[uzzi38]

So, this is what Intel is exploring?

Intel to use Nanowire/Nanoribbon Transistors in Volume ‘in Five Years’

www.anandtech.com

That is what everyone else calls GAAFET.

Everyone is researching it, not just Intel. Samsung are doing it at 3GAE and TSMC at 2nm.

 

[peonyu]

Getting tired of this joke. You do realize that even a room heater uses >1000 watts on high, right?

If you are so tired of it then just skip my post instead of hitting that reply button and posting lame comments like that, i bet you are fun at parties [not].

 

[LikeLinus]

 

[Zucker2k]

Linus tests a 10900k with coolers ranging from an Intel stock cooler to an AIO...

 

[mopardude87]

Linus tests a 10900k with coolers ranging from an Intel stock cooler to an AIO...

Thank you for this post! Been thinking about a 10900k a while for a dedicated gaming box, been kinda out of the loop on how bad off this chip is on cooling. Like people would talk about it, but few will post something as useful as this video. Looks like my DR4P will work, but its just the bare minimum its looking like.

Edit: Nvm