Monday, May 11th 2020

Intel Core i9-10900K Stressed, Package Power Reads 235W, Temperatures 93°C

A stock Intel Core i9-10900K 10-core processor was subjected to FPU stress by Chinese PC enthusiast @WolStame. The power and temperature values of the processor are inside HEDT territory. With a Furmark GPU stress running on the side, under AIDA64 FPU stress, the i9-10900K measured a package power draw of up to 235.17 W, as measured using HWInfo64. The CPU package temperature shot up to 93 °C. A 240 mm AIO liquid CPU cooling solution was used in the feat. Interestingly, the processor is able to sustain clock speeds of 4.77 GHz, which is close to the advertised 4.80 GHz all-core turbo boost frequency, called for by the multi-core FPU stress.

To show that the values weren't obtained in a few seconds of test, the AIDA64 Stability Test window keeps a timestamp log and displays time elapsed into the stress. In this particular case, the all-core stress has been running for close to 48 minutes; and yet the processor is keeping up with its advertised all-core boost speed, making this an impressive feat.
Sources: @WolStame (Weibo), @9550pro (Twitter)
Add your own comment

110 Comments on Intel Core i9-10900K Stressed, Package Power Reads 235W, Temperatures 93°C

#51
londiste
10900K at 4.8 GHz is nothing to sneeze at. It is likely to get pretty close to 3900X in terms of performance.
Posted on Reply
#52
Vya Domus
DR4G00N
Could also just be a hot running sample, not common but some chips just run super hot even with custom water and tame voltages.
"A hot running sample" might dissipate an additional 20-30W, chips don't vary that much.



The difference is the market is responding completely different to those two similar situations. Whereas AMD back then was universally considered behind from a technological point of view people still marvel how for the third time in a row Intel bolted two more cores to the same design watching how power goes downhill every time they do it.
Posted on Reply
#53
Houd.ini
jmcosta
This seems to be about right, if you stress test lets say a Ryzen2 FPU it will thermal throttle as well, even at stock settings with a good cooling.
the FPU subtest drives the power draw and thermal state of the system to the maximum regardless of whether it is overclocked or not


Where are these graphs from? I run folding@home on my 3700x, which uses AVX, and it does not get anywhere near these power consumption levels ( I am assuming the top graph is in watts, don't know whether it is CPU only or total system or whatever).
Posted on Reply
#54
EarthDog
Houd.ini
Where are these graphs from? I run folding@home on my 3700x, which uses AVX, and it does not get anywhere near these power consumption levels ( I am assuming the top graph is in watts, don't know whether it is CPU only or total system or whatever).
F@H is not AIDA64 FPU. ;)
Posted on Reply
#55
TheoneandonlyMrK
EarthDog
The difference though, power use and pricing aside, it is a competitive product when comparing core to core and single threaded performance. This isn't remotely like the steaming piles of bulldozer/vishera/piledriver was compared to its modern counterpart (and frankly, intel CPUs multiple generations behind).

Performance-wise, it is a good product. It just costs more and uses more power.

If we look at the chart that was posted earlier, there isn't much difference between it and a 9900k wattage wise.... and the 10900k has 25% more cores/threads.
It's also not performing that much better per core, would you recommend an upgrade path from a 9900k to this for a casual gamer?.
I hope not.

You would Have to be a pro or enthusiast for that.

And frankly you had to OC FX to 5.5Ghz to get anywhere near this wattage.
That much wattage can't so easily be glanced over IMHO.
Posted on Reply
#56
londiste
Vya Domus
The difference is the market is responding completely different to those two similar situations. Whereas AMD back then was universally considered behind from a technological point of view people still marvel how for the third time in a row Intel bolted two more cores to the same design watching how power goes downhill every time they do it.
FX9590 struggled to compete with contemporary i7 4770 in everything except a few well-threaded integer-heavy applications (largely thanks to CMT vs SMT).
i9 10900K is a bad comparison due to 10 cores but at 8 cores i7 10700K will likely be quote competitive vs R7 3700X/3800X at the same core count in most areas except power draw.
theoneandonlymrk
It's also not performing that much better per core, would you recommend an upgrade path from a 9900k to this for a casual gamer?
For a casual gamer all of that is overkill. If we add price into the equation, even more so. R5 3600 or upcoming i5 10400F is where it is at for a gamer. Now that Intel has SMT again, these two should be pretty much on par with what they can do - i5 a bit ahead in gaming, R5 a bit ahead in productivity.
Posted on Reply
#57
EarthDog
theoneandonlymrk
It's also not performing that much better per core, would you recommend an upgrade path from a 9900k to this for a casual gamer?.
I hope not.

You would Have to be a pro or enthusiast for that.

And frankly you had to OC FX to 5.5Ghz to get anywhere near this wattage.
That much wattage can't so easily be glanced over IMHO.
These arent all marketed towards a casual gamer, though. It would be silly to upgrade from ryzen 2000 to ryzen 3000 in the same manner (regardless if that gain is more, its still mostly negligible.

Maybe 5.5 ghz on ancient fx, but stock on 3900x seems to do the same thing, no? Did I read that chart incorrectly?
Posted on Reply
#58
TheoneandonlyMrK
EarthDog
These arent all marketed towards a casual gamer, though. It would be silly to upgrade from ryzen 2000 to ryzen 3000 in the same manner (regardless if that gain is more, its still mostly negligible.

Maybe 5.5 ghz on ancient fx, but stock on 3900x seems to do the same thing, no? Did I read that chart incorrectly?
That's with two more core's and in a specific load, it's the same enough though ISH.
I agree on the buyer, it's not typically last generation owners but older systems.
Many of which will even need a new PSU for those loads.

But I do take issue with the fact you wanted to put aside power use and price.

Power use, price, performance and use case are the driver's for change and upgrade and you seam to want to dismiss half of people's reasons to upgrade just to make these viable.

Power use and price are more important to some than others but never irrelevant.

Your point regarding F@home v Aida 64Fpu

Is noted, Intel are and have pushed (GPUwtaf Nokia) FPU technology beyond AMD but is not a typical use case.

Perhaps in using more reasonable and realistic daily loads the power use difference could be gauged better, I mean the likes of folding at home, WCG, and other fairly high load applications that more people will experience.

Damnit phones are shite at autocorrect.
Posted on Reply
#59
heky
EarthDog
F@H is not AIDA64 FPU. ;)
Sure, but i just did a 30 minute AIDA64 fpu run on my 3900x and the consumption is nowhere near 250W. At least looking at HWINFO64 or Ryzen Master. Its roughly @ 150W.
Posted on Reply
#60
EarthDog
theoneandonlymrk
Power use and price are more important to some than others but never irrelevant.
Actually, I'm not dismissing their relevance, but trying to share perspective. People are posting things like these processors are "shit"......but they aren't. The only thing out of line is power use. The price, though higher, doesn't seem to be "terrible/shit" IMO. Let's look...... :)

Keep in mind these sales/price drops are in part, a response to these CPUs hitting the market.

3900x - 12c/24t - SEP = $499($424 on sale at newegg)
3800x - 8c/16t - SEP = $399($339 on sale at newegg)
3600x - 6c/12t - SEP = $249($204 on sale at newegg)

i9-10900K - 10c/20t - SEP = $488
i7-10700K - 8c/16t - SEP = $374
i5-10600K - 6c/12t - SEP = $262

I also understand that prices at launch will likely be higher...yada yada.... but day 1 to day 1... outside of the flagship, core for core the Intel offerings are less expensive. It took months on the market and the imminent release of worthy competition to lower them. That said, what is on the webpage at the time is the decision maker and they are more expensive. However parity in pricing is starting to show (thanks AMD for Ryzen/competition!).
heky
Sure, but i just did a 30 minute AIDA64 fpu run on my 3900x and the consumption is nowhere near 250W. At least looking at HWINFO64 or Ryzen Master. Its roughly @ 150W.
Maybe that is system draw? No idea... I just see images and no link so I can't follow up.
Posted on Reply
#61
CrAsHnBuRnXp
MxPhenom 216
An AIO you mean with very little fluid and terrible fluid flow rates. A legit loop with a good, maybe double thick radiator 240 or 360 would probably be fine.
Most people arent going to be doing that though.
Posted on Reply
#62
svan71
jmcosta
This seems to be about right, if you stress test lets say a Ryzen2 FPU it will thermal throttle as well, even at stock settings with a good cooling.
the FPU subtest drives the power draw and thermal state of the system to the maximum regardless of whether it is overclocked or not


I wonder why the 3950x is excluded?
Posted on Reply
#63
Vayra86
EarthDog
Actually, I'm not dismissing their relevance, but trying to share perspective. People are posting things like these processors are "shit"......but they aren't. The only thing out of line is power use. The price, though higher, doesn't seem to be "terrible/shit" IMO. Let's look...... :)

Keep in mind these sales/price drops are in part, a response to these CPUs hitting the market.

3900x - 12c/24t - SEP = $499($424 on sale at newegg)
3800x - 8c/16t - SEP = $399($339 on sale at newegg)
3600x - 6c/12t - SEP = $249($204 on sale at newegg)

i9-10900K - 10c/20t - SEP = $488
i7-10700K - 8c/16t - SEP = $374
i5-10600K - 6c/12t - SEP = $262

I also understand that prices at launch will likely be higher...yada yada.... but day 1 to day 1... outside of the flagship, core for core the Intel offerings are less expensive. It took months on the market and the imminent release of worthy competition to lower them. That said, what is on the webpage at the time is the decision maker and they are more expensive. However parity in pricing is starting to show (thanks AMD for Ryzen/competition!).

Maybe that is system draw? No idea... I just see images and no link so I can't follow up.
While true, the per core price difference is not quite enough to offset the power requirement gap, total cost of ownership might turn out very similar or in the advantage of AMD. So from a cost perspective, I'm not so sure the argument holds. What you dó get, at a similar price, may be single core performance in some minor ways. Very minor at this point, but fact remains Intel still does turbo a lot higher. It doesn't always pay off though compared to Zen.

It becomes a different story when you also factor in the cost of cooling. Then the price argument leans heavily in AMD"s favor.
Posted on Reply
#64
ERazer
MxPhenom 216
An AIO you mean with very little fluid and terrible fluid flow rates. A legit loop with a good, maybe double thick radiator 240 or 360 would probably be fine.
CrAsHnBuRnXp
Most people arent going to be doing that though.
exactly most ppl re-use their old stuff, your implying 240 aio that was good enough for 9900k now need custom loop?
Posted on Reply
#65
TheoneandonlyMrK
EarthDog
Actually, I'm not dismissing their relevance, but trying to share perspective. People are posting things like these processors are "shit"......but they aren't. The only thing out of line is power use. The price, though higher, doesn't seem to be "terrible/shit" IMO. Let's look...... :)

Keep in mind these sales/price drops are in part, a response to these CPUs hitting the market.

3900x - 12c/24t - SEP = $499($424 on sale at newegg)
3800x - 8c/16t - SEP = $399($339 on sale at newegg)
3600x - 6c/12t - SEP = $249($204 on sale at newegg)

i9-10900K - 10c/20t - SEP = $488
i7-10700K - 8c/16t - SEP = $374
i5-10600K - 6c/12t - SEP = $262

I also understand that prices at launch will likely be higher...yada yada.... but day 1 to day 1... outside of the flagship, core for core the Intel offerings are less expensive. It took months on the market and the imminent release of worthy competition to lower them. That said, what is on the webpage at the time is the decision maker and they are more expensive. However parity in pricing is starting to show (thanks AMD for Ryzen/competition!).

Maybe that is system draw? No idea... I just see images and no link so I can't follow up.
Well I am not saying that ,never did and fair enough, pricing will help/matter.

But I did say we need better more contemporary chart's of power use that align better with user's typical use cases not just FPU power draw, which show's these in a bad light on per core power use.

Show power use with game's running, blender, WCG ,you probably can name a few yourself.

I'm not suggesting we just use cinebench either note.
Posted on Reply
#66
Vayra86
theoneandonlymrk
Well I am not saying that ,never did and fair enough, pricing will help/matter.

But I did say we need better more contemporary chart's of power use that align better with user's typical use cases not just FPU power draw, which show's these in a bad light on per core power use.

Show power use with game's running, blender, WCG ,you probably can name a few yourself.

I'm not suggesting we just use cinebench either note.
A much more interesting test these days I think is limiting CPUs at a TDP budget that is doable with different kinds of cooling.

Then compare the actual performance. Because right now the bigger differentiator for K-CPUs and Ryzen is the way you use them and limit them in each use case. A lot of potential buyers are no longer interested in the balls to the wall performance level because it will cost a lot to get there.

A writing on the wall is that we see more and more topics about undervolting these days, than overclocking...
Posted on Reply
#68
EarthDog
Vayra86
While true, the per core price difference is not quite enough to offset the power requirement gap, total cost of ownership might turn out very similar or in the advantage of AMD. So from a cost perspective, I'm not so sure the argument holds. What you dó get, at a similar price, may be single core performance in some minor ways. Very minor at this point, but fact remains Intel still does turbo a lot higher. It doesn't always pay off though compared to Zen.

It becomes a different story when you also factor in the cost of cooling. Then the price argument leans heavily in AMD"s favor.
I'm not worried about the distributed platform peeps (F@H/BOINC, etc)... that is an entirely different use case and I would agree with your sentiment. If there is a difference of 100W, for 8 hours a day, for a year, with my electricity cost, that is a couple of bucks. Of course that varies.

You get comparable performance all around. Most things you do are single threaded and need clocks, not width and threads.

Intel is NOT the primary choice here, I agree overall, but the haters hating calling it shit and overpriced clearly need some perspective. Cooling can be different, but is there really a problem when a 240 AIO is able to keep 235W in check? People aren't using 240mm(+) AIOs on 3900x to keep it cool too?? FPU in A64 is also a worst case scenario...
theoneandonlymrk
Well I am not saying that ,never did and fair enough, pricing will help/matter.

But I did say we need better more contemporary chart's of power use that align better with user's typical use cases not just FPU power draw, which show's these in a bad light on per core power use.

Show power use with game's running, blender, WCG ,you probably can name a few yourself.

I'm not suggesting we just use cinebench either note.
Is that supposed to be to me, the rest of this post past the first sentence? I'm not talking any of that jazz either... just trying to put some perspective on things as a few have really missed the mark.
Posted on Reply
#69
Vayra86
EarthDog
I'm not worried about the distributed platform peeps (F@H/BOINC, etc)... that is an entirely different use case and I would agree with your sentiment. If there is a difference of 100W, for 8 hours a day, for a year, with my electricity cost, that is a couple of bucks. Of course that varies.

You get comparable performance all around. Most things you do are single threaded and need clocks, not width and threads.

Intel is NOT the primary choice here, I agree overall, but the haters hating calling it shit and overpriced clearly need some perspective.

Is that supposed to be to me, the rest of this post past the first sentence? I'm not talking any of that jazz either... just trying to put some perspective on things as a few have really missed the mark.
Agreed, its not as grim as some make it out to be. Me included, but that's also just for giggles :p
Posted on Reply
#70
TheoneandonlyMrK
Vayra86
A much more interesting test these days I think is limiting CPUs at a TDP budget that is doable with different kinds of cooling.

Then compare the actual performance. Because right now the bigger differentiator for K-CPUs and Ryzen is the way you use them and limit them in each use case. A lot of potential buyers are no longer interested in the balls to the wall performance level because it will cost a lot to get there.
As user's of Ryzen we both know how that would go and it sure wouldn't be in Intel's favour , it's easy to limit Ryzen to 65watts via bios switch on some boards and yet it doesn't effect boost clock's as much as you expect.

No I feel it isn't that reasonable to limit CPU Tests like that since out of the box performance is more relevant to consumers personally.
Posted on Reply
#71
Vayra86
theoneandonlymrk
As user's of Ryzen we both know how that would go and it sure wouldn't be in Intel's favour , it's easy to limit Ryzen to 65watts via bios switch on some boards and yet it doesn't effect boost clock's as much as you expect.

No I feel it isn't that reasonable to limit CPU Tests like that since out of the box performance is more relevant to consumers personally.
Note the 'different kinds of cooling'. Out of the box for me would be on a dual stack air cooler. For another it might be custom water.

I still believe the goal of reviews should be to inform potential buyers. Not a pissing contest that just shows us maximum performance that is highly conditional.
Posted on Reply
#72
londiste
theoneandonlymrk
As user's of Ryzen we both know how that would go and it sure wouldn't be in Intel's favour , it's easy to limit Ryzen to 65watts via bios switch on some boards and yet it doesn't effect boost clock's as much as you expect.
You would be surprised. My AM4 board used to have a setting for cTDP that went away with BIOS updates, came back but didn't work and then went away completely. AGESA/Ryzen Master settings did not work for power limit.
Posted on Reply
#73
Wet_Paint
jmcosta
This seems to be about right, if you stress test lets say a Ryzen2 FPU it will thermal throttle as well, even at stock settings with a good cooling.
the FPU subtest drives the power draw and thermal state of the system to the maximum regardless of whether it is overclocked or not


The power consumption for 9900K seems waaaaaaay too high.
Here's what I get with my 9900K at 5.1Ghz (at max LLC), memory is running at 4270Mhz which probably contributes to higher power consumption too.
Posted on Reply
#74
EarthDog
I heard that with a 360 AIO and i9-10900K at 5.2 ghz all c/t ~1.35V with Aida's default l test it passed at 100C. TJmax on these are 115C, note.
Posted on Reply
#75
kapone32
The numbers seem to match with the strength of the VRMs on Z490 boards.
Posted on Reply
Add your own comment