try changing to active (available since Kernel 6.3) and set the default behaviour to power or balance_power - power stands for powersaving here opposed to performance on the other side of the scale. Details on how to do that can be found here:
with that, and AMD ECO Mode activated in BIOS, I get very good temperature readings on my 7700X that are only above 80° when compiling stuff with a Alpenföhn Brocken 3 air cooler - no other config changes made.
Edit: I totally forgot that I also apply a dynamic undervolting on all cores with the AMD Curve Optimizer in BIOS. The values one can reach there heavily depend on the individual hardware lottery, and one could get more out of it when applying per core values, but I don’t have the time and all core -32 works fine here. (64 is max)
CPU Usage usually indicates total usage across all cores while temperature is usually highest single core temperature so high temps while seemingly “low” CPU usage is normal. Most of the replies here are for power saving when idle / low activity but you get high temps while gaming so I don’t think these suggestions will help. You can look into limiting FPS, lowering down game settings or setting fans to maximum.
What I meant is that you can get high temps even if your CPU usage is %10 or %100 because all it takes is a single core to go full blast and reach high temps.
That depends on the game setting. If you do not limit FPS and enable ultra setting (it requires a lot of calculations by CPU and GPU), that would cause high temperature when you play for a long time.
That is by design and should be a normal behaviour. As far as I was able to find. There is no target frequency when CPU goes to boost mode but there is a target thermal limit ~85˚C.
I have Ryzen 9 7900 and it does the same thing. Well, actually it doesn’t since it hits 85W power limit before the temperature can go any higher (I like to keep the power consumption low which also keeps the cooler quiet). Interestingly when all 12 cores (24 threads) run at 100% load the temperature is around 65˚C at 4.8GHz but when only 2 cores run at full load the temperature is 80˚C at 5.4GHz. Also, undervolting helped significantly to drop power consumption and increased clock frequency.
I found several articles talking about how AMD tells reaching 95º its “normal temps” (which I totally disagree with)
The thing was: I swap my CPU and decide “i’m gonna run a linux on this” so i was convinced it was my fault messing up something, and don’t find anything about this until yesterday, when a friend told me about that and change the focus to the hardware.
Reaserching a bit more drive me to some guides on how to limit the TDP and set a power target into the bios to a reasonable 85ºC, + changing the curve optimizer to -15.
Here is a how to video on an ASUS bios Fix to Ryzen 7000 high temps
I also do not feel like that is normal or good but I could not find a strong evidence why is it a problem - just a feeling. Somewhere, I read that this is exactly what graphic cards do and they are working just fine.
As far as I can tell the high temperature has an effect on the power consumption and perhaps a “wear” of the cpu (I upgrade abpout every 6 years so I hope it will be fine for so long).
And the cpu is so far doing just fine. There is no thermal throttling - the frequency of the cpu is stable at the higher temperature and there is no drop in a performance over time.
Since I locked my cpu to 85 W max power and I have more than sufficient cooler the temperatures are reasonable between 60-80˚C at 100% load (depending cores used and clock speed). The tepmerature is very closely related to the power consumption (and performance). You can limit the power but it may influence the performance as well.
A good way to reduce the power consumption (and temperature) with the same performance is undervolting - if your cpu has a good quality silicone.
