• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

CPUID - Hardware Monitor

Joined
Jul 18, 2007
Messages
813 (0.12/day)
Location
Almonte, Canada
System Name Sonny Boy
Processor i5 11600K 4.9GHz @ 1.35 Vcore
Motherboard AORUS Z590 Elite AX
Cooling Deepcool Gammix 240mm AIO
Memory 4 X 8GB Teamgroup DDR4 3200MHz
Video Card(s) ASUS RTX 3060 Ti GDDR6X
Storage Adata Legend 1TB
Display(s) Philips 27" 4K
Power Supply Corsair 750W
Mouse Razer Basilisk
Keyboard Razer Huntsman V2
Software Windows 11 Pro
For some time I have been using Speedfan, Coretemp, RealTemp, and ASUS Probe to measure my CPU temperatures. Since Speedfan, Coretemp, and RealTemp all seemed to agree on precisely the same core temperatures I assumed everything was rosey. ASUS Probe doesn't report core temperatures... just the "CPU" temperature.

Well my "CPU" temperature as reported in my BIOS, Speedfan, and ASUS Probe has always been about 8C hotter than my cores which is ridiculous since nothing should be hotter than the silicon itself. So I just assumed that the CPU temperature was messed up.

Now I am starting to question this. Let me explain...

I have always trusted CPUZ as an excellent tool to display my CPU and memory information. I noticed that they (CPUID) had a new tool called CPUID Hardware Monitor so I thought I would give it a whirl. What I discovered was that it displayed the "CPU" temperature the same as BIOS/Speedfan/ASUSProbe but it had a whole new take on the Core temperatures. In fact it made sense if the CPU temp was supposed to be Tcase since the cores were then about 7C hotter than the case.

So now I am asking myself... I have been assuming all along that the core temperatures were correct and the "CPU" temp was screwed up... Could it be that the CPU temp is indeed correct and it is the core temps that are off?

See my screenshot below. Anyone know how CPUID is getting their core temps and why it is so different from the other tools?

:confused:
 

Attachments

  • CPUID_HWM_E2160_380X9_1.39V_fans_60.jpg
    CPUID_HWM_E2160_380X9_1.39V_fans_60.jpg
    164.5 KB · Views: 1,397
cputin (+ speedfan cpu temp) is from the sensor chip, horribly inaccurate

the temps under e2160 are read from the internal temp register just like coretemp and realtemp. realtemp uses some additional calibration formula iirc
 
cputin (+ speedfan cpu temp) is from the sensor chip, horribly inaccurate

the temps under e2160 are read from the internal temp register just like coretemp and realtemp. realtemp uses some additional calibration formula iirc

OK. So my initial assumption that my core temperatures as read by Speedfan, Coretemp, and Realtemp are correct and useful is correct. And I like to see my cores <60C although Tj max is ~85C. So at full load I am <50C which = :)

Only question remaining is how is CPUID getting their core temperatures? Seems they are calculating them rather than reading them from the on-die sensors. Or they are reading them and applying a "correction" factor. Either way it seems that it is 3 against 1 :cool:
 
Only question remaining is how is CPUID getting their core temperatures? Seems they are calculating them rather than reading them from the on-die sensors. Or they are reading them and applying a "correction" factor. Either way it seems that it is 3 against 1 :cool:

CPUID hardware monitor is using Tjmax = 100 so it shows your cores as 15C hotter then the other programs. 47 + 15 = 62.
 
CPUID hardware monitor is using Tjmax = 100 so it shows your cores as 15C hotter then the other programs. 47 + 15 = 62.

In that case they should just be reporting a "distance to Tj max" as RealTemp does. Reporting it as an actual temperature would be misleading since just increasing Tjmax does not change the actual silicon temperature... just the headroom until it melts :)

Thanks
 
Back
Top