9900K on air - too hot?

i don't think it works that way, i mean even Tom's Alphacool Ice Age 2000 chiller (1500W cooling capacity) could only cool theirs down to 63°C.



https://www.anandtech.com/show/13400...600k-review/22

https://www.tomshardware.com/reviews...u,5847-12.html


Paul seems to agree that air cooling an OCed 9900K is asking a bit too much.

Hmm, this is convincing me that I might need to re-do my cooler paste. Pook is only hitting 90C on 200W and Tom's Hardware is hitting 90C on 229W with a similar cooler (H110i AIO). I won't be setting any records on air but seems like I should be able to do at least 5C better with what I have.

The Tom's article actually has really good data, but I don't really agree with the assessment that you can't run the 9900k on air. I'm not having any problems running stock clocks on air, it just gets a bit limiting when trying to push the overclocks.

Sheeeet man friend of mines rig I upgraded with a 9600k its a pig on voltage compared to some 9900's. He has a Le Grand Macho with Phobya Nanogrease good airflow and at 5ghz its hitting mid 80's peaking in 90 range OCCT cpu test, (forget about AVX).

The only way I bring it in control is to use the boards AI overclock with a custom cooler point and optimism scale and took me forever to keep it from randomly going up to 5.1 ghz I got fed up with trying to overclock it manually. I think hes back to stock to be honest for longevity reasons.

Transistor density baby.

well its not like Tom's is implying that air cooling is impossible, its just that it would be stressing both the cooler and chip.
e.g. expect it to hit 70°C or higher even at stock is what hes implying.

by the way, is your macho RT on it's stock fans? have you thought of swapping them with TWO PWM 2000RPM fans?
in normal workloads the temperature should be much lower, so the fans should remain at low RPM, only when its on really high load would it rev all the way up.

Yes, RT is on the stock fan. It's extremely quiet in this case. It starts getting audible once you hit about 70% speed, but only if the case fans are at idle (which they won't be if the CPU is hot enough to push the RT fan over 70%). Basically there's no rational combination of CPU and case fan speeds where you'd actually be able to hear the CPU fan over the case fans.

What 2000 RPM fans did you have in mind? How much do you think that would impact core temps?

https://noctua.at/en/nf-a14-industri.../specification
This fan, or if you want to go all out then try the 3000? though i wouldn't recommend it.
https://noctua.at/en/nf-a14-industri.../specification

of course it would impact the temps, at the very least the sheer airflow is much higher.

Well, with my LGM RT, I'm in the mid 70s with a 23-24C intake temp with my 1.32v 4.5ghz 1680 v2 (also works out to be around 200w load in realbench). Add in a much higher thermal density, and a thicker(?) die, it's no surprise it's running hot on air. There are some things you can do to help with airflow, but chances are it's not going to knock off 10C+ unless you run some fast spinners in there or get wet.

definitely not a massive 10°C+, but knocking it back down below 90°C should be possible.
his temps shows 91°C average, 96°C peak.

What temps do you get with realbench stress test?

Also, it could be worth your time to take off the cooler and inspect the cold plate along with the mounting screws. If the base of the cooler is convex, then that could be a variable for high temps as it won't make as much contact as a flat one (fyi, mine is pretty flat). You also have the mounting hardware not being torqued in there good enough which can cause a loose fit. Paste tim does degrade over time, but not by much, and re-applying annually is fine.

He has an Asus Maximus XI. His vcore sensor is the VR VOUT. Elmor mentioned this.