• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

7950x overheating

D

davfel01

Joined
Aug 3, 2021
Messages
36 (0.03/day)
here is what I found.
I have a 7950x AMD and Asus x670e-e gaming motherboard. Shuts down only when uefi bios setting AI overclock set to optimized.
If I run OCCT benchmark it shuts the system down. powers off.
So i investigate and I find just 1 core that is going about 15 degrees C higher than all of the other cores.
my conclusion is that it is NOT cpu cooling, I use a waterkill IV waterblock.
It is impossible for just one core to run hotter than all of the other cores and be the cooler or thermal paste.
So, I am returning the CPU and replacing with new.
For what its worth, other than the cpu shutting the power off in OCCT benchmark it seem to run fine.

Also. it is easy to check by running cinebench23 with CPUIDHardwareMonitor open and looking at Core(max) Temperature by opening the +(plus) sign in front of core(max).
you will see the temp of all cores and if 1 core is 15 or 20 degrees higher I would replace the chip. Maybe several cores are running hot. They should NOT run at a significant difference.

I will let you know after the new CPU comes in.
Dave
 

Attachments

  • Screenshot (9).png
    Screenshot (9).png
    1.5 MB · Views: 381
Last edited:
quite strange. I've at least never heard of such a problem before, maybe someone else has and would know what caused it.
 
I think many of the complaints about power offs and reboot were caused by this. People did not investigate far enough. Just my opinion
 
ir_cow said:
Who didn't? Its your computer lol
Click to expand...
I think he was saying that about other people's computers, but also yeah I've never heard anything about anyone complaining about shut-offs before that didn't have to do with the psu or cooler.
 
It's probably a bad sensor. Happens to me years ago with a 8350. Ended up getting a RMA and problem solved. It never shut down though, just high temps on one core.
 
Sensor only reads high under heavy load. Not medium or light. At lower loads all of the sensors are in the same ball park.
 
davfel01 said:
Shuts down only when uefi bios setting AI overclock set to optimized
Click to expand...

Did you also test temps with AI overclock disabled? It's extremely unlikely that a Ryzen CPU will shut down from thermal issues Derbeur was able to boot into windows with no cooler at all installed. Ryzen CPUs can adjust clocks every 1 ms so they are very good at throttling performance in thermally adverse environments. It's not impossible to be a defective CPU although I would definitely turn off the AI overclock and any other OC settings (excluding EXPO / XMP, although you should run MemTest if you haven't already).
 
davfel01 said:
Yes. It does not power system off but still core 10 is high
Click to expand...

Try HWInfo and see if that reports the same thing. Best to double check IMO.
 
davfel01 said:
Hwinfo does not have the same reading that I could find
Click to expand...

By that do you mean it has different temperature values or that you couldn't find the same values?

When you start up the program make sure you click sensors only, the per core temp readings will look like this:

1681179998829.png
 
Gigabyte Rx6500 xt gpu.
Segotep 750w gold psu

Occt program has a nice feature of telling you last know state before crash.
It shows core 10 at 116 degrees c.
The rest not bove 90 c.
I am going with bad processor. Possibly core or sensor. I will know in a day or 2.
 
davfel01 said:
Gigabyte Rx6500 xt gpu.
Segotep 750w gold psu

Occt program has a nice feature of telling you last know state before crash.
It shows core 10 at 116 degrees c.
The rest not bove 90 c.
I am going with bad processor. Possibly core or sensor. I will know in a day or 2.
Click to expand...
well that's not quite what I expected to be with a Ryzen 9 7950x, are you using this as a workstation? also are you reusing an old cpu cooler?
 
Personal computer. I use a heatkiller IV waterblock

Again as I said previously, not likely a cooler problem, all cores would read high. And they do not.
 
I will go with faulty software readings, I've seen 100C+ readings in hwinfo64 too which can't possibly be right, these CPUs have a tjmax of 95C. Hwinfo64 has an entry which tells you if the CPU is in fact thermal throttling, unless this is triggered I would not worry :

1681198448221.png


I have a core (core 5) on my 7900 that runs much hotter than the rest as well, I've found that it's mostly because windows schedules everything on this core first so it's utilized more than the rest.
 
Sorry, I disagree. The system can go over 95 c. When it does it throttles back supposedly. I do have to worry because my system powers off
 
davfel01 said:
I do have to worry because my system powers off
Click to expand...
What I am saying is this is probably unrelated to this, it could just be instability, is PBO on ?
 
Ai optimized on system shuts down. Turn it off system runs. That core is still quite hot

These CPUs handle tjmax differently. They go right 95 and do their thing. It is not limited to 95
 
davfel01 said:
Ai optimized on system shuts down. Turn it off system runs. That core is still quite hot
Click to expand...
I don't know what that AI optimized thing is, is that some auto OC feature ? I am talking about the regular PBO setting in BIOS.

davfel01 said:
These CPUs handle tjmax differently. They go right 95 and do their thing. It is not limited to 95
Click to expand...
Like someone else said you can run these CPU without a cooler and they would still not shut down, just throttle, I am pretty sure something else is causing your system to shut down. If you are convinced it's the CPU RMA it I guess.
 
I believe and temp limit is 115. My system gets to that when running occt benchmark. It shuts off and after rebooting and open occt it show last known states and it shows that core going to 116

I do not see any other problems except for this. I will know for sure in a day or 2
 
The reason I am saying all of this is because I saw 120C once in hwinfo64 and the system didn't shut off, it was obviously an incorrect reading in my case.
 
You must log in or register to reply here.
Back
Top