How high is too high for CPU temps when I'm gaming? What can be done about high operating temps?

I have a 5 year old desktop gaming computer with an AMD Ryzen 5 5600X 6-Core processor. It has the colored fan lights, so I set it in my Aura software for the light to be temperature-controlled-- anything below 50C is green, 50-80 is yellow, and anything over 80C is red. So I can easily and quickly see what range my CPU temp is at.

I usually almost never see the red light come on during gaming, unless my fan intake screens have gotten dusty. A quick dusting and I’m solidly back in the yellow zone (or alternating green to yellow if I’m not gaming). But I just started playing Cyberpunk 2077, and I’m noticing it pegs my CPU up into the red zone fairly regularly. Not all the time, and it doesn’t last too long. But still a concern.

I’ve heard AMD processors are made to run a bit hot, and looking it up, Google’s AI tells me:

The absolute maximum safe operating temperature (Tjmax) specified by AMD for the Ryzen 5 5600X is 95°C (203°F). While the processor will not instantly damage itself at this limit, it will automatically slow down (thermal throttle) to prevent overheating, and you should ideally keep sustained loads well below this threshold.

So, are occasional brief spikes over 80C acceptable? Can I just set my upper limit for the red zone to 85C and call it a day? I know I can also adjust settings down like FPS within the game, and I will probably do so.

I would not be worried at all about those temperatures. The occasional spike above 80°C will not harm your CPU in any way. It really isn’t until you get up to TDP max that you have to worry, and even then, the CPU will throttle to protect itself.

Now, it may not be running as cool as it could: beyond cleaning the intake, you could consider trying to repaste your CPU. But I wouldn’t be worried.

I would see no issue with changing the threshhold temp for your red light to 85 or even 90°C and seeing if it stays yellow.

I’m going to disagree with the estimable @BigT here - 80c is my personal redline in my tech support practice, agreeing with the Aura software mentioned.

Does this ONLY happen with your CPU, never your GPU? Because IME if the CPU is hotter than the GPU there is something wrong. First check all the fans are working, especially the ones in the CPU housing. If that is okay you likely have a thermal paste problem, see about getting it renewed.

Wow, two very different replies! First, I appreciate and value both replies, and despite how different they are, I don’t think either one is completely wrong or completely right.

@BigT’s reply, beyond simply being the most reassuring answer, is in line with AMD’s operating temperature recommendations, which are higher than for other CPU brands. I am aware of this, but a max of 95C still seems very high to me-- that’s over 200F. Seems pretty damn hot.

@Askance , to answer your questions, I haven’t been tracking my GPU temps. The Aura software ‘smart’ setting just says it tracks ‘temperature’ which I assume only means my CPU, though I have separate software that can track my GPU temps-- I’ll check what they peg out at next time I try to play Cyberpunk.

My fans seem to be working as they should. A reapplication of thermal paste seems like it couldn’t hurt to do.

Oh, and as for 80C being the ‘red zone’ in the Aura software, that was my setting, not Aura’s. It was recommended in a thread I started for advice on this computer that for AMD I could do a range of 50-80C for the yellow zone. So, theoretically at least, I could go 85C for my red zone setting and still be 10C under the operating limit.

If it isn’t throttling and the fan noise isn’t too loud, the temp doesn’t really matter. People get giant coolers so they can push more power through the CPU and run it even faster or because they want the PC to be very quiet (or both, if you’re willing to spend a lot.)

It could be a sign of potential problems though, if the temperatures don’t make sense for the workload. I don’t think 80~ degrees in 2077 is a sign of that though, especially with a stock or other lower-tier cooler.

Also you should set your lights to turn red based on CPU load. That seem much cooler to me. You want your PC to be as cool as possible, right?

Thanks, Palooka, appreciate the input. Is it possible to get aftermarket cooling solutions that I can add on my current box, I wonder, or would I need to get a whole new custom tower?

As for the RGB lights, I think I’d rather go with ‘useful’ than ‘cool’ (this more or less applies to my daily life, as well :smirk:).

Besides, wouldn’t CPU load and temperature variance be more or less in lockstep-- load goes up, temp goes up?

It might be tough to find an aftermarket cooler at a price you like, assuming you still have that G509 with the narrow Lian Li case.

The CPU cooler clearance is only 135mm, which puts it into small form factor territory. That category is now dominated by AIO water coolers but you don’t have compatibility for those.

There are SFF models with fan configurations parallel with the motherboard. One of those is your best bet, but they have a premium on them. There are cheap smaller tower coolers with small fans on them too, but that may not be a large improvement over the AMD stock cooler. Stock coolers used to be absolute garbage and those are a hold over from that era. “Low profile” is the search keyword.

You’re running into the limitations of a lower tier case. It isn’t causing problems for your PC to be clear, but it limits your options.

A lot of good tech answers here. I’ve found some simple non-tech methodology that helps. My gaming machine is not on the floor or even on the bottom rung that is like 4 inches above the floor. I keep it on top of my computer desk. Also, I keep my desk a foot from the wall so that the exhaust fans have ample air space and the heat doesn’t build up. I also have a small fan that I run behind my monitor that keeps the air moving away from the exhaust. Little things like that have definitely had me running cooler.

I’ve been replaying Baldur’s Gate 3 and noticed it hits my CPU fairly hard. Not quite to the point of throttling but close to 95C sometimes and not much below 85C the rest of the time. It’s one of the only programs that causes this so I know it’s not a thermal paste or heatsink issue and eventually just set some lower power limits on my CPU. In my case, I used Intel’s “Extreme Tuning Utility” (ironic that I’m tuning it downward, so extreme!) but I assume AMD has some similar solution that doesn’t require going into the BIOS.

I asked Claude for some limit suggestions – worst that happens is it sucks and I revert them – and it’s worked great. I’m now capping at 85C which is not only less worrisome but also less heat being pumped into the office this summer. No noticeable loss in performance. Results would be CPU dependent since some just run hotter or less efficiently than others but it might be worth looking into.

Thanks @Jasmine . I do have my PC on a tabletop away from any obstructions, but putting a small fan in the back to help facilitate air outflow is a good low-tech thing to try.

And thank you @Jophiel - I’ll look into if there’s an AMD solution for fine-tuning settings with the CPU.

My thinking is along the lines of what is safe and what you need to do. If I’m building a system or doing any upgrades, then I agree that staying well under 80°C is better. My point was just that it’s not unsafe or going to reduce your longevity if it sometimes goes above that.

@solost The equivalent AMD software is called Ryzen Master.

Thanks!

So, a couple updates if anyone’s interested-- Google’s AI gave me some suggestions on how to limit some game graphics in CP 2077 and to set the max frame rate in my NVIDIA app to 60FPS. So yesterday I downloaded the Ryzen Master app, and without changing any CPU tuning yet, I noticed that, with those game changes, I never got back into the red zone playing CP 2077 with my RGB lights with the Aura app set to show red at 80C.

I did notice that there’s a slight temp difference between what the Ryzen Master app says the temp is, getting up to the low 80s a few times, and what my Aura RGB light settings showed, which should have triggered red at those temps, but didn’t. I noticed the same on the low end-- when the RGB lights go green, meaning the CPU should be at 50C or lower, the Ryzen app would still say low 50s. But it’s just a few degree diff, and staying below 85C according to the Ryzen app I considered a win. I didn’t really notice much if any falloff in game graphics quality.

Then tonight I tried setting the CPU to ‘Eco Mode’ in the Ryzen Master app, which I read online should do a great job setting the CPU to run significantly more efficient and cooler with little cost in gaming performance. And, playing the game, it did seem to have a significant cooling effect on the CPU while I was playing-- staying mostly in the upper 70s C, quite a bit better than mostly low 80s last night. Did the game seem to stutter a bit while playing? Maybe? Just my imagination?

I also monitored my GPU temp while I was playing, using a different app called Open Hardware Monitor; I never did that before while gaming. It was also running pretty high while playing, upper 70s to low 80s C, topping out at 83 at the very highest briefly. It’s at a cool 37C right now. Google’s AI tells me:

While video gaming, an NVIDIA GeForce RTX 3060 normally operates between 65°C and 80°C. Temperatures up to 83°C–85°C remain within acceptable safety thresholds depending on your specific card model and case airflow, while NVIDIA sets the maximum safe limit at 93°C.

…so I guess my GPU is operating just within acceptable limits.

I think I will try the thermal paste reapplication on the CPU once I get a chance to pick some up. If that makes a big difference in lowering temps, I’ll remove some of the game limits I imposed.

Keep it up and one day you’ll turn into the kind of weirdo who has a special sensor monitor on his desk to always know what the temps and usage are. I’ve heard those people exist, anyway.

The Ryzen Master app is probably getting the direct CPU temp, using the hottest core, while your RGB settings are getting some sort of system temp or averaging things out. There are a bunch of different temperature sensors.

Given what I see the GTX 3060 running in benchmarks, a 60fps cap can make sense to smooth out the rendering.

If you are seeing any slowdown on the CPU, the settings I’m seeing that are at least somewhat CPU dependent are Cascaded shadow range, Level of Detail, and Crowd Density, as well as Distant shadow resolution and Max dynamic decals. The shadow ones are least likely to be noticeable, as they basically affect stuff in the distance. Level of Detail should as well.

(I knew about Crowd Density, but I looked up a few more based on some optimization guides, rather than trusting the AI search results directly.)

Heh, doesn’t seem weird at all for a dedicated gamer. If you’re putting a sports car through its paces, you’re going to be keeping a close eye on the dashboard tach, engine temp, etc., right?

In fact, I got a birthday coming up. Got an idea for a birthday gift to myself now… :thinking:

Some of those settings you mentioned overlap with recommendations by the AI, like Crowd Density; others the AI recommended were things like cloud / steam formation. Some of the settings you mention are different, so if I continue to have CPU concerns maybe I’ll try lowering those too. Thanks!

This is not a good rule of thumb at all. I have a geforce 4080 super that almost never goes above 70c and CPUs routinely exceed that in normal operations. It really depends on what CPU, what GPU, and the cooling solutions for each.

To the OP: It’s probably not going to hurt anything to run your CPU relatively hot if it’s below the thermal throttle temp, which sounds like it’s 95c. You’d generally prefer cooler than hotter, but the consequences are usually pretty low if it’s not hitting that limit, especially if it’s only occasionally hitting it.

I would recommend the freeware utility HWInfo64 for your temp monitors. It has a ton of data and different sensors and records high low and mean for as long as you leave it open. And importantly for your purposes, it has a thermal throttling monitor that will tell you if any thermal throttling has occurred. At least it does on my CPU, I would assume it does on your model of CPU too but I’m not sure it’s universal. If it never hits thermal throttling, I wouldn’t worry about it too much.

Sounds like you already dust out your case. Make sure the fans blow air in directions that make sense - generally intake is from the bottom and front and exhaust is from the top and back. There should be a logical air flow path where the air comes in cold, runs over your components, and exits the case. It should be designed like this by default, but every once in a while there’s a funky setup or a fan blowing in the wrong direction. In general, the 5600x isn’t a super hot CPU, but a small form factor case does reduce ventilation in general.

Run HWinfo64, see if there’s thermal throttling, and if there isn’t, and it’s not routinely running up against the edge, I wouldn’t worry about it.

Thanks SB, great advice! I’ll look into getting HWInfo64.

Update: so, making adjustments to the Cyberpunk 2077 specs, capping frame rate at 60, and setting the CPU to ‘eco mode’ seems to have made the CPU happy, so I had put off buying thermal paste and doing a reapply to the CPU for now.

But lately, playing CP2077, I’ve been noticing some glitchy stuff like stuttering and brief freezing, especially during battles, which is of course doubleplusungood. And sometimes the road just glitches out, so I’m driving on nothing. A consequence of my game graphics downgrades, just a glitchy game, or something else?

So I decided to monitor my GPU temps again-- last time I checked it was getting hot during play, but I thought within acceptable limits. But last night, whenever I’d get into a battle in CP2077, the temp would spike up to mid 80s C and stay there, even after the battle, until I’d do something like open my inventory-- then it’d cool down to like low 70s. So I took a cursory look at the GEFORCE RTX 3060 GPU card: it’s mounted low in the box, with only about an inch or so clearance between the fan underneath and the bottom of the box. I blew some canned air and got some dust out from the fan blade, but it didn’t help much, and I can’t really get to the fan to clean it well without removing the graphics card.

So, looks like I’m going to have to get me some thermal paste, and, per this video, take the GPU apart, clean it thoroughly and re-paste it (and of course the CPU too while I’m at it). If that doesn’t help much, I guess I’ll get MSI Afterburner and tweak the GPU settings-- jack up fan speed under load and such. Under low to normal GPU loads, like right now, it’s cool as a cucumber-- 33C at the moment.