i9 12900k temperature problem

[ir_cow]

This thread is still going? Once again me and @Vayra86 (among others) have pointed out time and time again. Alder Lake temps depend on the workload, voltage and P1=P2.

If the MB has this TDP "unlocked" aka 4093 Watts, a 12900K will pull 260-270 watts in Cinebench. One run isnt enough to thermally soak a cooler....

If you have it locked at the AIO setting or Air cooler setting, it will be less. The % loss is like 9% from 150 TDP to 270 as wiz covered in one of the articles ( and posted here). I use Alder Lake every day for reviewing. Don't tell me a air-cooler is capable of handling long durations of heavy loads with a 12900K. Unless that P1=P2 has been changed, its total BS.

 

[fevgatos]

This thread is still going? Once again me and @Vayra86 (among others) have pointed out time and time again. Alder Lake temps depend on the workload, voltage and P1=P2.

If the MB has this TDP "unlocked" aka 4093 Watts, a 12900K will pull 260-270 watts in Cinebench. One run isnt enough to thermally soak a cooler....

If you have it locked at the AIO setting or Air cooler setting, it will be less. The % loss is like 9% from 150 TDP to 270 as wiz covered in one of the articles ( and posted here). I use Alder Lake every day for reviewing. Don't tell me a air-cooler is capable of handling long durations of heavy loads with a 12900K. Unless that P1=P2 has been changed, its total BS.

Once i go home ill prove it. Default settings on my apex, 4096 watts power limit, 10minutes cbr23. Is that test okay with you?

It certainly doesn't draw 270w on cbr23, I get that much oced to 5.2+ ghz

The 270 watts you are quoting are for system power, it says right there on the graph. It does not consume more than the stock 240w whether you have the power limits removed or not. Actually with power limits removed it seems to consume 15-20w less, which aligns with my findings. Mine runs at 225w at stock

So you might be using it everyday for reviewing but your numbers are just wrong

Here you go, it doesn't get any more stock than that, everything was reset to default from the bios, PL1=PL2=4095w. It does NOT consume 270 watts in cbr23. You say you use it everyday for reviewing, I call BS. You should have known how much it consumes if you did. There is no freaking way a stock 12900k hits 270w at stock even with removed PL's in cinebench, lol.

 

[ir_cow]

I cannot comment on R23 yet, but R20 does. Every single Z690 MB I reviewed pulls 270+ in R20. I'll give it a go tonight and see. I have a 12600 and 12900K to try it with.

I have verified this the best I can with multi-meter and amp clamp. The software is accurate (within 5%)

Once again I stand by my reasoning. you're not pulling 270w your air cooler is soaking up 200 watt which is acceptable for a air cooler. Now if you're still trying to say that your air cooler as good as a AIO or water cooling you're wrong because your only at 200 watts.

Why your 12900K is only drawing 210 in R23, IDK. Maybe its that software, I'll find out soon enough.

 

[fevgatos]

I can overvolt it a bit and make it pull 240w. It will still handle it like a charm. As I've said, it starts throttling at about 280 to 300 watts.

 

[tox1c90]

Your claim is simply not true. If I remove power limits (=4096W) on my MSI Z690 Tomahawk which is actually the default with this board, my 12900k draws 290W average in CB23 (all voltages on auto) reported by hwinfo (package power).

If it consumes much less on your build, then it’s probably your board applying less voltage.

 

[fevgatos]

Your claim is simply not true. If I remove power limits (=4096W) on my MSI Z690 Tomahawk which is actually the default with this board, my 12900k draws 290W average in CB23 (all voltages on auto) reported by hwinfo (package power).

If it consumes much less on your build, then it’s probably your board applying less voltage.

My claim also aligns with techpowerups claim, he reports 270w SYSTEM consumption on cbr23. What's your voltage under cbr23? Mine is 1.19

 

[ir_cow]

You

My claim also aligns with techpowerups claim, he reports 270w SYSTEM consumption on cbr23. What's your voltage under cbr23? Mine is 1.19

You may claim 270 but your screenshot is 210..

1.15-1.2v is stock voltage from turbo from what I've seen. I suspect your power limiter is in fact enabled in the BIOS.

 

[fevgatos]

You

You may claim 270 but your screenshot is 210..

1.15-1.2v is stock voltage from turbo from what I've seen. I suspect your power limiter is in fact enabled in the BIOS.

You know the difference between system and CPU right?

No the power limiter isn't enabled, you can actually see it in hwinfo. It lists power limits. Its right there on the screenshot

 

[ir_cow]

You know the difference between system and CPU right?

This is getting old... CPU Package Power : Peak 216. Sorry for rounding down before.
What are you trying to defend again? Please explain to the popcorn audience, I'm sure they are at a loss as well.

 

[fevgatos]

This is getting old... CPU Package Power : Peak 216. Sorry for rounding down before.
What are you trying to defend again? Please explain to the popcorn audience, I'm sure they are at a loss as well.

That your first post at the top of this page is wrong. 12900k does not pull 270w on cinebench according to me and techpowrups review. You are also wrong about air coolijg not being able to keep it from throttling. I have a small single tower 120 cooler and it does the job brilliantly.

 

[Alessa]

Once again me and @Vayra86 (among others) have pointed out time and time again. Alder Lake temps depend on the workload, voltage and P1=P2.

I have CPU temperature spikes also with the 12900k but i am not sure from where they come from..
They do not seem to come from workload or Vcore voltage or Watts used..

using 50ms refresh on HWinfo i am trying to find out what is causing these temperature spikes without any luck..

The temperature spikes suddenly hit + 80 C without any apparent big spike in core usage, or Vcore voltage, or Cpu Package power, or disk write activity....


Any idea how i can find out where the CPU temperatures spikes comes from?

maybe by looking at other graph value in HWinfo while the temperature spikes?

 

[eidairaman1]

Is it crashing?

 

[P4-630]

Or get a 46W TDP Celeron.

 

[Alessa]

Is it crashing?

It does not seem to be not crashing.

Is there a way to pinpoint the cause of the 12900k temperature spikes using HWinfo..?

maybe i should monitor some other HWinfo value graphs..?

 

[ir_cow]

@Alessa Why is your vcore at 1.37~ V?

 

[Alessa]

@Alessa Why is your vcore at 1.37~ V?

it is the voltage that the Asus z690 extreme AI overclocking kinda choose for my AI all core 5.3ghz overclocking, with small E-core disable, also according to the Load line 4 setting i choose to reduce the temperatures.

it stays at 1.37v because i am using Windows 7,
and for some reason the boost seems to be broken and all core are always at 5.3 ghz .

 

[P4-630]

I remember my Asus Z170 board also used excessive core voltage when set to "auto overclock", I think it's better to test and set Vcore and overclock manually if you want to overclock.

 

[Alessa]

@Alessa Why is your vcore at 1.37~ V?

on this recorded 73C temperature spike the Vcore voltage seem to have dropped to 1.296V...

Maybe it was that that raised the temperature?


could it be crashing like eidairaman1 said?

Is it crashing?

 

[eidairaman1]

on this recorded 73C temperature spike the Vcore voltage seem to have dropped to 1.296V...

Maybe it was that that raised the temperature?


could it be crashing like eidairaman1 said?

Crashing is lockups/freezing or BSODs

 

[Alessa]

Crashing is lockups/freezing or BSODs

it does not freeze or BSOD, but these temperature spikes seems abnormal.

Many other people also have temperature spikes with the 12900k
maybe it comes from the same problem as mine...
it would be great if we could find a way to pinpoint the cause of these spikes..

If HWinfo sensors with 50 ms refresh cannot find the problem,
should i try using the ROG True Voltician oscilloscope that comes with the Asus extreme motherboard..?

 

[ir_cow]

Looking at the graph you provided, the spikes happen during a write session. So the system is doing something. I am also not surprised by those temps at 1.35+ V. If you did a real stress test, it would be in the 90s or hit the 100 Tj.Max.

 

[eidairaman1]

Looking at the graph you provided, the spikes happen during a write session. So the system is doing something. I am also not surprised by those temps at 1.35+ V. If you did a real stress test, it would be in the 90s or hit the 100 Tj.Max.

The OS does maintenance in the background while idle, it does not take coffee breaks like it did during xp days. The OS auto Trims/Defrags, to disable it requires a registry setting

To me the OPs "problem" is not a problem but the natural order of things established by software and hardware makers.

The only way to know for sure is to run the system bare bones with the OS and Drivers, no rgb software, no codecs.

No games/loaders.

This is why I manage services.msc and msconfig, task scheduler

Also WUD has a nasty habit of phoning home like malware/virus busting tools.

 

[Athlonite]

Something is running in the background use task manager to keep an eye on what processes are using CPU time when idle

JayzTwoCents did a handy dandy undervolt video for the 12900K give it a watch

 

[Daftfader]

These new 12000's serise react diferant to older gens, esecially when all the auto OC bells and whisltes are turned on. Spikes on single cores are 100 normal as that core gets picked to process something.

Now if over time ALL you temps magically get hotter, or the clocks magically get more OCed. That would be the AI cooler grader not reacting fast enough possibly. If I had my PC on idel for ages, my coolers effeciency score would go up, and then if randomly after days of it being barely taxed, I turned on hitman3 with settings maxed, my acceptable clocks at max load would of crept up so much i used to imsta crash (my mobo manufacture asus) has fixed it so its not as agressive as at first, but there's many AI sliding scales effecting these CPU's so things randomly changing over time is not uncommon. If OCing, a mild one of using AI with no AI controlling your cooler efficiency (but imp any auto ai OC on these CPUs is just a waste of power apart from one rarely used) as it OCs the most when your not using it, and throttles when you are.

It coule be its inabiloty to change its settings after a period of low use thats caused temps to suddenly go up, amd may go back down over time given u are using it a lot.

Try by resetting your cooler ai, running timespy for 30 mins, then immidiatly going and setting your cooler AI to trained and to never update your cooler score. This caused my temps and clocks to creep up over time, doesn't sound like your exact issue, but hopefully give you an idea of where to start and what these CPU's mobos are doing behind the scenes. A loooot of AI controlled OC related stuff. And if you exact CPU in on the boaderline of two presets it may be switching depending on load or time ised etcetc.

Idk, hopefully it helps you work out the issue!