Conflicting reports for CPU temperatures for Speedfan & Bios

New to PC silencing? Read & post your questions here. Dedicated to rosy_toes.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Bluemilk
Posts: 23
Joined: Wed Mar 16, 2011 9:12 am

Conflicting reports for CPU temperatures for Speedfan & Bios

Post by Bluemilk » Mon Mar 21, 2011 8:22 am

I was using Speedfan to monitor my newly-built PC, and these are the (worrying?) results. (see attached image).

http://img543.imageshack.us/i/speedfan.jpg/

Which of the temperatures actually represent the actual CPU temps?
In the bios and the coretemp program, the temperatures closely match the Core 0 temperatures.
Is it anything to worry about?

The motherboard I am using is the Asus P67-LE B3 board, and cooler is the Noctua NH-U12P SE2. Can I begin my over-clocking adventures or do I need to reseat the cooler?

lodestar
Posts: 1683
Joined: Fri Aug 05, 2005 3:29 am
Location: UK

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by lodestar » Mon Mar 21, 2011 9:19 am

There's definitely something wrong with that CPU temperature reading. The CPU will have a generic CPU temperature sensor somewhere and then there will be Digital Thermal Sensors on each core. BIOS CPU temp readings will be from the generic temperature sensor. Since the only significant source of CPU heat is the cores I would normally expect that the reported CPU temperature would be less than the individual core temps reported. I have tried SpeedFan 4.43. Compared to CPUID HWMonitor Pro all the readings agree, apart from the core temps. SpeedFan is reporting the core temps as 5C less than HWMonitor Pro, but CoreTemp reports the same readings as HWMonitor Pro. So I am inclined to believe that SpeedFan is not reliable for Intel CPU core temps.

Bluemilk
Posts: 23
Joined: Wed Mar 16, 2011 9:12 am

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by Bluemilk » Mon Mar 21, 2011 5:15 pm

lodestar wrote:There's definitely something wrong with that CPU temperature reading. The CPU will have a generic CPU temperature sensor somewhere and then there will be Digital Thermal Sensors on each core. BIOS CPU temp readings will be from the generic temperature sensor. Since the only significant source of CPU heat is the cores I would normally expect that the reported CPU temperature would be less than the individual core temps reported. I have tried SpeedFan 4.43. Compared to CPUID HWMonitor Pro all the readings agree, apart from the core temps. SpeedFan is reporting the core temps as 5C less than HWMonitor Pro, but CoreTemp reports the same readings as HWMonitor Pro. So I am inclined to believe that SpeedFan is not reliable for Intel CPU core temps.
What does the CPU temperature measurement in Speedfan actually measure? Also, what is this Aux measurement? Is it too high?

lodestar
Posts: 1683
Joined: Fri Aug 05, 2005 3:29 am
Location: UK

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by lodestar » Mon Mar 21, 2011 10:18 pm

It might be worth downloading CPU HWMonitor, which is a free download, from http://www.cpuid.com/softwares/hwmonitor.html and see what that reports.

Bluemilk
Posts: 23
Joined: Wed Mar 16, 2011 9:12 am

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by Bluemilk » Tue Mar 22, 2011 12:08 am

lodestar wrote:It might be worth downloading CPU HWMonitor, which is a free download, from http://www.cpuid.com/softwares/hwmonitor.html and see what that reports.
This is the report from CPUID.

http://img834.imageshack.us/i/cpuidp.jpg/

quest_for_silence
Posts: 5275
Joined: Wed Jun 13, 2007 10:12 am
Location: ITALY

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by quest_for_silence » Tue Mar 22, 2011 12:21 am

Bluemilk wrote:Can I begin my over-clocking adventures or do I need to reseat the cooler?

Are the SpeedFan and CPUID temps taken at load? And at which fan speed? I would give a check also with the popular Real Temp 3.60 (or 3.65), usually to be preferred for Intel CPUs.

At anyway, if I were you I would however re-seat the cooler, but before cleaning the thermal paste, I would also take two photos of the heatspreader and of the cooler base (to let us see whether the paste footprint look good or not).

Bluemilk
Posts: 23
Joined: Wed Mar 16, 2011 9:12 am

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by Bluemilk » Tue Mar 22, 2011 12:36 am

quest_for_silence wrote:
Bluemilk wrote:Can I begin my over-clocking adventures or do I need to reseat the cooler?

Are the SpeedFan and CPUID temps taken at load? And at which fan speed? I would give a check also with the popular Real Temp 3.60 (or 3.65), usually to be preferred for Intel CPUs.

At anyway, if I were you I would however re-seat the cooler, but before cleaning the thermal paste, I would also take two photos of the heatspreader and of the cooler base (to let us see whether the paste footprint look good or not).
Those were taken at idle.

The temperatures from realtemp are attached.
http://img15.imageshack.us/i/realtemp.jpg/

quest_for_silence
Posts: 5275
Joined: Wed Jun 13, 2007 10:12 am
Location: ITALY

Re: Conflicting reports for CPU temperatures for Speedfan &

Post by quest_for_silence » Tue Mar 22, 2011 1:09 am

It would look like you have no problem at all: nor with your temps, neither with SpeedFan.
To be sure, give a run with 4 instances of Prime95, you should not exceed the average 55-60°C per core (maybe no more than one core could go as high as 60-65°C).

The "CoreN" readings look like virtually identical among the three tools (for Core0 - the only one always visible in your schreenshots - I read Speedfan 34°C, HW Monitor 36°C, Real Temp 34°C).
That "CPU" sensor is probably a motherboard sensor (possibly placed in the wrong place), and so you have probably just to exclude it from Speedfan monitoring.

Post Reply