Showing results for 
Search instead for 
Did you mean: 


Journeyman III

Ryzen 5900x temp is 60 °C at 10-40% load.

So idk if it has to be like that. My 5900x is staying 60 °C even if it's 10% loaded. Whatever this load type is. When playing games it's loaded about 40% and stays 60 °C. When it's 100% loaded it's 70 °C. Does it have to be like that or not? Because i can't even watch a movie without noises from my PC.

17 Replies
Adept III

It can be confusing on these very high core count CPUs. A 10% load in task manager on a 5900X means 2 logical cores are 100% pegged and 1 core is 40% utilized. And 40% usage in task manager means 9.6 logical cores are pegged.

Zen 3 is a very small process, so that heat is concentrated in a tiny area, and depending on your cooling, 60C is not at all unreasonable for 10% load on a 5900X.

If your fans are spinning up too frequently and it's annoying, I would suggest adjusting the fan curve so it doesn't do that, and takes more time to both spin up and spin down.

I have EK-AIO 360. So you want my fans to spin like 20% on 60 degrees and 80% at 70 degress or what? smh


Do you have your cooler set as intake or exhaust? What case do you have, and also what is your GPU? I noticed that before I transplanted cases and my AIO was set as a top exhaust and zero rpm mode on the 6900xt it would increase the air temp inside the case by 10-15C and my cpu would idle near 60-65. After the transplant with the AIO set as the side intake and zero RPM mode turned off on the gpu my cpu is pretty firm at 43-45 under light use and hardly ever gets up to 70 under a full load. I think the advice given for setting your fan curve is probably pretty solid, keeping a slightly higher base fan speed will help with the spikiness of the temps and fan ramp ups. Also it may be a good idea to look at the curve optimizer in the PBO options on your motherboard and doing a slight undervolt. I have mine set at a negative 20 mv offset while also having the scaler set to 10x and a 200mhz boost offset on.



also while im not even touching the pc (laying on the bed for example) fan spins may randomly ramp up for like no reason. is that ok for ryzen?


This is correct, you must adjust fan curves as you like. For example I have fans on 800rpm and curve flat to 80C then increasining speed from 80C 800rpm to 90C 1700rpm

Moreover for watercooling only logical radiator fan control is control with respect to water terperature, everything else does not make sense.


I wouldn't say it's OK because it's annoying as hell, but it happens on my 5950X also.


If you say that it's annoying as hell, it means that result of assembly and personalization - meaning result of your actions and doings - of your PC is annoying as hell. It is you who set fan profiles and customizes your PC.

Other thing would be if it is not possible to customize fan curves of your cooling products, but it is not fault of 5950x, but particular cooling products.

Or am I missing something important here?


sounds totally normal for an AIO with a fan curve adjusted for an Aircooler.

Mine did that right after assembly too - adjusted the fan curve and it stays quiet and behaves properly. 

Temp. Spikes are totally normal for Zen3


Yes, you are missing something. I can load-test my 5950X all day long with intensive single-thread small FFT AVX benchmarks and never go over 67C, but then randomly when browsing it hits >75C and the fans spin up. This is a problem with the CPU or AGESA.

Exactly, the temp spikes are "normal" in that they happen for pretty much everybody. But it really shouldn't do that.


@JoltCola wrote:

but then randomly when browsing it hits >75C and the fans spin up. This is a problem with the CPU or AGESA.

No it is no problem, it works correctly, burst CPU loads create spikes and as I wrote only problem is how you personalized your build, how you set your fan curves. Also you do not know what caused performance burst, maybe some short term intensive workload, antivirus, etc. which is normal. Do not expect fix to AGESA, bios, motherboard, os, everything works correctly.

Nevertheless I can provide info how I personalized my PC, how I solved this "issue" with fan settings in following picture. I have flat curve until 80C and it ramps up from 80C to 90C 100%, this is how



Incorrect. Random background processes can't cause as much load as a power virus burn-in tool. This only happens on AMD, not Intel.


They can. You live in the past, where Intel with old process dominated. AMD is on new process with higher heat density. This is the way. Get used to it.

You can read article e.g. from Noctua for 7nm process for Zen 2



Again, you are wrong. This isn't about overall heat generation, or how heat is concentrated in a smaller area. This is about random background spikes that generate brief spikes much hotter than running an actual power virus.

If the spikes were to 67C, in effect roughly the same as that power virus, that would be reasonable. 10C hotter is not and cannot be legitimately hand-waved away.

Something is going on when a nearly idle ryzen core wakes from sleep with substantial load. Perhaps it's given too much voltage for some reason-- I do have LLC active, but only at level 2. Something to investigate, anyway.


You do not know what load was in background. You must run e.g. hwinfo for some time where temperature spikes are and look at power load. You claim idle but you do not know if there was running something intensive or not.

Run hwinfo and post screenshot of all sensor values, it will show that there is some load in your system, some bad software - e.g. rgb utilities, various manufacturer tools, etc.


First computer I built myself was a Pentium-90. I've been doing this awhile now, I'm not new, so please spare the basic troubleshooting. I have been running hwinfo64 24/7 for the past several weeks, since I was debugging incredibly annoying WHEA reboots, later fixed in AGESA It is indeed a sudden load spike causing greatly increased temps for a moment. My theory is LLC may explain why these very brief spikes generate more heat than running a power virus for an hour.


Can you share screenshot of all hwinfo sensor data running in time in which temp spike occured?


to spare basic troubleshoots  - so far i built 24 Zen3 Systems - everyone of those has Temp Spikes.

For the fan ramp use a fancurve suitable for an AIO;

for reference i only used NZXT X73 AIOs here atm since i had them in stock - all core load with prime/cinebench or all core "power virus" - not that an all core load wouldnt already hit each set default limit i get around 62-63°C with the rad on top as exhaust.

< then all core loads i get spikes up to 72°C with different games. 

Idle at the Desktop i have around 33°C at 21°C room temp - that changes with Windows Updates for example -> bam 63° down to 55 then up again (spikey) till those are finished.

Talked to several ppl so far and -> its normal for Zen3... do what you will with that information or ignore it - not gonna dabble in the who built more systems since the stone age :) - This is the Way