Question CPU temp randomly jumped to 100-105 C!

[lhall6788]

Hi All, Wondering for some feedback if possible please.

Last night I booted up my PC, which is a beast might I add.

Specs are:

AMD 5900x
X570 F tuf mobo
RTX 3080
H115i Plat water cooler AIO
32gb 3600hz ram
M2 1tb x 2
etc

I have core temp running all the time, i had the PC on for 3? minutes and noticed the temp showing on Coretemp was 103'c at first site. I HOPED coretemp was having a moment, but i am informed that the temps shown are straight from the bios so it is likely to have hit those temps.

During those 3 minutes, I would have booted up COD Vanguard beta and closed it (due to the sheer amount of updates) and I jumped into discord.

I rebooted, and it was back to it's usual Idle 43'c and 82'c at Peak playing COD on Max settings.

(This is the first time it's done it and I have had the PC for 6 months, might be worth knowing)

I did update Armory Crate and iCUE in the afternoon as some of the LED's weren't syncing. I turned PC off at 6pm, jumped back on at 8pm.

I've now set Overheat protection on Coretemp, just incase the PC doesnt do it via. BIOS (which i thought the BIOS would auto shut down if it hit those temps, i checked bios but I can't see a setting to set the temp for this).

My main point and question on this is:

During those 3 ish minutes, would that have done any damage to the CPU or Thermal Paste?

I'm an over thinker and I now find myself constantly staring at coretemp... which is very annoying, but it's a £3k PC...

My friend seems to think it would either be a temp air block in the water cooler (however the fans and lights were all operational at the time of seeing temp peak) OR Software not booting up the pump on startup.

Any feedback would be greatly appreciated.

Thank you
Lee

 

[dorsai]

Fans and lights working doesn't mean the AIO pump is working...is it making any weird sounds ?

How is your AIO cooler mounted ?

View: https://www.youtube.com/watch?v=BbGomv195sk

 

[MeanMachine41]

Yes you do have a beast so best to check for stability for piece of mind.
Are or have you attempted an Overclock on the CPU?

Some temperature apps read with differing sensors.
The best is t/die sensor for Ryzen and HWinfo64 has that so check it out.

There are synthetic stress testers that can test thermals and one I use and recommend is Aida64. It's free and will test your CPU, FPU, Cache, RAM, and GPU.

Run the test for 20mins and take screen shots at the 20min mark of relevance to your Voltages and temperatures which you can have analyzed here. If you want an analysis then upload your desktop images to IMGUR and obtain the bburl as a link here.

 

[NightHawkRMX]

Make sure you removed the plastic from the bottom of the aio.

 

[lhall6788]

Fans and lights working doesn't mean the AIO pump is working...is it making any weird sounds ?

How is your AIO cooler mounted ?

View: https://www.youtube.com/watch?v=BbGomv195sk

No strange sounds, Cooler is mounted on the top of case, PC is a 3XS Build from Scan and was fully stress tested before shipping etc.

Assume the BIOS controls the pump starting up? can it 'misfire' , or equiv and fail to start?

Thank you

 

[lhall6788]

Make sure you removed the plastic from the bottom of the aio.

Lol, thanks, I can only hope Scan remembered this when building!

 

[lhall6788]

Yes you do have a beast so best to check for stability for piece of mind.
Are or have you attempted an Overclock on the CPU?

Some temperature apps read with differing sensors.
The best is t/die sensor for Ryzen and HWinfo64 has that so check it out.

There are synthetic stress testers that can test thermals and one I use and recommend is Aida64. It's free and will test your CPU, FPU, Cache, RAM, and GPU.

Run the test for 20mins and take screen shots at the 20min mark of relevance to your Voltages and temperatures which you can have analyzed here. If you want an analysis then upload your desktop images to IMGUR and obtain the bburl as a link here.

Thank you

As mentioned, it's the first time in 6 months that it's done this, and it's not done it since.
Happened pretty much minutes after booting up.

I'll give HWinfo64 a look now, but as mentioned in another reply, this is a 3XS build from Scan and they fully test the build prior to shipping and replace anything that is faulty.

Appreciate your reply

 

[lhall6788]

My main worry was thermal damage to the CPU and Thermal Paste really.

Looking at the CPU info on HWinfo64, it says that the CPU Thermal Trip Limit is 115.0 °C & CPU HTC Temperature Limit is 115.5 °C, so I would assume for the few minutes that it was running between 100-105'c, it wouldn't have done too much/any damage? what do you think?

Appreciate the replies, thanks all.

 

[mrhoot1]

I've got the exact same thing happening to me with hwinfo64.

 

[scout_03]

to test the pump for any air leak in open sysyem start hwinfo and press one of the pump tubing to see if it made any change in temp and return after you stop pressing it .

 

[drea.drechsler]

My main worry was thermal damage to the CPU and Thermal Paste really.

Looking at the CPU info on HWinfo64, it says that the CPU Thermal Trip Limit is 115.0 °C & CPU HTC Temperature Limit is 115.5 °C, so I would assume for the few minutes that it was running between 100-105'c, it wouldn't have done too much/any damage? what do you think?

Appreciate the replies, thanks all.

If it hits 115 or 115.5 I think it will shut down immediately, like tripping a circuit breaker would.

TJmax is 90C any temp above that and it throttles itself to bring temp below it. I don't believe it was above 90C for more than a few milliseconds at any one time before the throttling brought it back under.

NO monitoring program works that fast it's always showing something that happened maybe two seconds previous and somehow integrates that reading with the next one it polls which could be two more seconds. If it jumps up and back down as it lowers clocks you never see the true picture.

It's normal and expected for 5900's to push to 90C, but hitting above 100C isn't supposed to happen.

 

[Karadjgne]

Most monitoring programs will poll the temps every 250-500-1000ms. Which you can't read. What you'd see would be digital '88' looking numbers. So they are deliberately either averaged (Ryzen master) over like 3 seconds, or written at the 3 second mark etc.

Unless you graph the polls, temps you see are subjective. If you had a 3 second report on a 1000ms read, you could have 100, 100, 70, 90...etc and you'd see a 70°C temp, not the 100°C spikes. Conversely you could have 70, 70, 100, 70, 70...etc and for 3 more seconds all you'd see is the 100°C spike temp reported.

That applies to All software that polls/reports, Ryzen master averages the 3 seconds, so you'd see 89° for the first and 80° for the second, both of which can be misleading because of the dip in load or the spike, but generally accurate as to the overall working range of the cpu.

The cpu will see the actual temps as polled, and respond as it needs to but isn't going to throttle over a temporary spike.