cpu core temps too high on i7

[mikewse]

Running the latest hwinfo64 on a Toshiba laptop equipped with an Intel i7-4500U cpu, it seems to consistently show cpu core temps that are roughly 15 degrees Celsius too high.
The room has an ambient temp of 25 degrees, and on initial startup speedfan reports core temps of 25, while hwinfo reports temps in the 38-40 area.
I have attached report and debug file.

 

[Martin]

And why do you think HWiNFO is wrong ?
It's not possible for the CPU when running to have the internal (core) temperature equal to ambient temperature. It would have to be totally turned off for quite a while. Even a slight workload produces heat that causes the internal temperatures to rise.
I have also checked the data you attached and the temperatures reported (in the 50 C range) are correct.

 

[mikewse]

Hi Martin,

I realize that ambient temp will not last for long ;-)

This is what I see and the way my reasoning goes:
1) I let the system cool down for several hours in sleep mode with hwinfo and speedfan already open.
2) When waking the system I get readings within a few seconds, and then hwinfo reports ~40 degrees and speedfan reports ~25 degrees. I don't have an exact measurement of the ambient temp, but it's somewhere in the 22-25 range.
3) After being powered on for a while in "idle", the readings are now hwinfo 50 and speedfan 35.

It seems logical to me that the temp doesn't climb to 40 in just a few seconds but I'm no authority in this area so I'm most certainly wrong. Anyway, the difference between hwinfo and speedfan is interesting so if you think speedfan is the one at fault I'll post a bug report there instead.

Best regards
Mike

 

[Martin]

Well, you might not believe it, but putting load on the CPU can cause the temperature to go up by several degrees (or even tens of degrees) in a fraction of second !
Also, the cooling needs some time to kick in and reduce the temperatures, which is a much slower process than the rise in temperature.
I'm not sure what SpeedFan reports as CPU temperature, maybe it's something else than the CPU core temperature, or it uses some correction factor.
According to the data I have analyzed I believe HWiNFO is correct. You might try other tools like CoreTemp etc to see who is right

 

[mikewse]

You're right. Coretemp shows the same as hwinfo so it's two against one. I'll report to speedfan.
I'm really sorry to have wasted your time :-(

 

[Stilez]

You're right. Coretemp shows the same as hwinfo so it's two against one. I'll report to speedfan.
I'm really sorry to have wasted your time :-(

For information, I would agree, Speedfan's temps have been dubious (for me anyhow) for a long time - my first note to myself on that was over 2 years ago. So it sounds likely. Hope this helps reassure you.