How come my CPU hits 90° on stress tests?
Posted by Sufficient-Ad-7349@reddit | buildapc | View on Reddit | 23 comments
My friends all say their CPu's just have the manufacturer fans and they work fine.
I do have an old watercooling system i can install but i'm confused about why this one wouldn't work.
keithstonee@reddit
What's wrong with 90 degrees?
No_Spare1827@reddit
what kind of stress test are we talkin? these CPUs are designed to go until they meet a power limit, a thermal limit, or a bottleneck somewhere else in the system. so if u are going a like actual stress test designed to put the CPU under max load then its doing what its supposed to. now if its a game or something that this is happening then u might have a problem.
is this PC older or are u just reusing the cooler
Sufficient-Ad-7349@reddit (OP)
Pasting my response to another person here:
"Hm, yeah, I tested due to crashes. Possibly due to my power cable being a bit loose, but my cat was also sitting on the pc so I started testing for heat issues because I don't think that alone should have been enough. I was only playing Zero Hour.
I asked here because the stress test flagged my cpu as a running way too hot, seemingly.
I ran a stability test in OCCT on my AMD Ryzen 5 7600 and immediately hit 90° before progressing to 95°. I'm seeing some thermal throttling but it's not awful.
Once I stopped the test, the cpu quickly returned to 60°
Idk if I actually have a game that will push this cpu this hard"
As for the cooler, it's the manufacturer's cooler. I had a really old PC, but I pretty much replaced everything that touches the motherboard with modern components. Most of the case fans are from the old build otherwise.
The old build did have a CPU water cooler (trying to make up for a shit CPU at the time) that I believe is compatible with my current build, it's just a matter of me finding enough justification to get off my ass and install it lol.
semidegenerate@reddit
Is 60C where it normally idles with no load? That sounds a bit high for a 7600.
Sufficient-Ad-7349@reddit (OP)
Not sure, i'll have to check when I'm home again. Thanks for pointing that out
SuspiciousArt7316@reddit
You either have a thermal paste issue, or case airflow issue.
Some newer GPUs like 9070xt have a vent that blasts hot air upward, which could then compromise the cold air intake for your CPU.
What case and GPU do you have? What type of fan setup in the case?
Sufficient-Ad-7349@reddit (OP)
Hm I have a 5060 but I forget what model. It's usually at about 30°
Forgot the case as well. It opens up in the top and on one side. I have three fans in on that side and one fan pushing out upwards (that one's not powered tho). I think i did the thermal paste flawlessly. Just an X as recommended. maybe a tad too much but i doubt it.
SuspiciousArt7316@reddit
Post a pic, can’t tell by your description.
Sufficient-Ad-7349@reddit (OP)
Ir's an Antec AX90.
Just got home and checked the cpu. It averages around 42° but 60° when I have several chrome tabs running
beirch@reddit
It's not warm, not sure what the other guy is talking about. Idling at 40-50C is totally normal for AM5 CPUs, and so is 60C with a light load.
You can follow this guide if you're stressing out about the temps you're getting under stress tests/multi thread work: https://youtu.be/FaOYYHNGlLs?is=Py4zlUFo4MhBLwG2
It should lower your max temps by a bit.
SuspiciousArt7316@reddit
That’s a tad warm.
With that case make sure you have the 3 front fans running. You don’t need the rear exhaust but it doesn’t hurt.
samkoLoL@reddit
if the temps also return to lower values fast, the problem isnt conductivity. you also said no games push the cpu that high, so you can just leave it as it is if you dont notice stutters caused by cpu throttling itself, or you can get a bigger cooler which will solve this issue, since you also said you only have stock cooler installed, which most likely just isnt that much material.
hextanerf@reddit
because they're supposed to
The-Copilot@reddit
What CPU? It's normal some while under extreme load to push itself to the thermal limit then throttle back to not exceed the max temperature. It's giving you It's max performance. I think all the newest Gen Intel and AMD CPUs do this.
If your PC was doing a thermal shutdown or hitting 90C while gaming, that's a different story.
You should also do one of the benchmarks that let's you compare to similar hardware. Don't expect to he top of the leaderboards, some people have crazy overclocks and cooling but if you are in the range then you are good.
Sufficient-Ad-7349@reddit (OP)
Hm, yeah, I tested due to crashes. Possibly due to my power cable being a bit loose, but my cat was also sitting on the pc so I started testing for heat issues.
I asked here because the stress test flagged my cpu as a running way too hot, seemingly.
I ran a stability test in OCCT on my AMD Ryzen 5 7600 and immediately hit 90° before progressing to 95°. I'm seeing some thermal throttling
The-Copilot@reddit
That is exactly what should happen during a stress test. The test is seeing how far you can push your CPU, so it should hit the thermal limit which is 95C on the 7600.
Basically it goes full "acceleration" until it hits 90C, then it slowly speeds up until it hits 95C and stabilizes the speed to maintain that max temperature. It gives you the max performance while staying right at the thermal limit. This was discussed by AMD during the release of AM5 and is fully intended.
Given your crashes, I would recommend updating your drivers like chipset drivers and also bios if you haven't already. Make sure you get drivers from the motherboard company's website. Windows auto installs drivers that are good enough to get the system running and they mostly work but they are meant to make the system work enough to install the proper drivers.
The AM5 platform being new meant it had some kinks that needed to be worked out and those updates including the bios increases the stability. There was plenty of complaints during release which is totally normal for a new new platform release and they have been cleared up since.
You can also check your event viewer to get more info on what caused the crash.
I got a 7900x during release and was working IT (hardware diagnostics and repair) at the time, so I've dealt with it a bunch.
Sufficient-Ad-7349@reddit (OP)
Thank you! This is very helpful
Baddger_223@reddit
Perfectly Normal when doing a stress test
CorrectEducation8842@reddit
90° during stress tests honestly isn’t unusual on modern CPUs, especially with stock coolers. Stress tests push way harder than normal gaming.
forevertired1982@reddit
If you run cinebench on the last 4+ generations of intel it will hot throttle temps very quickly even with a 360mm AIO a 14900k for example will boost until it hits throttle temps and then slow down to keep just beelow the throttle temps .
Amd 7000 is designed to hit thermal throttle temps or max boost clocks whichever comes first.
cdojs98@reddit
It's... a stress test... what were you expecting, exactly?
The only hardware understressed by such software is only doing so because the software has not caught up to it yet (probably due to sales volume such as 9950X3D or 14900K), but as Moore's law continues to hold mostly consistent, the software will eventually reach a point where it does stress that extant hardware to it's limits as well.
The speed at which you hit Thermal Throttle, or the amount of Compute Power lost at the point of Throttling - these are more relevant variables to manage. If it's taking 3, 4, or 10 hours to hit 90°C on stress testing, that's nothing to worry about - no game pushes the limits that much like a stress test. If you're hitting Throttle within the first 30 seconds of the test and losing a ton of Compute Power/Speed, then you should take that as a sign of inadequate cooling and address it with better hardware.
Live-Juggernaut-221@reddit
For many CPUs 90 is a perfectly normal temp
Naerven@reddit
What sort of test and what sort of hardware are we talking about?