HWMonitor will usually show about 5C higher than Coretemp since it is using
Tcase instead of Tjunction ( I think I got that right)
All that means is it uses a different way to factor temps
Also Realtemp will also be off because of different TJMax numbers it uses
Really at idle due to the slope of signal of DTS there is a lot of error
Your idle is more of a guideline than a exact measurement
whether your at 25c or 35c if that is normal operating idle with system
running optimally then what you dont want to see
is a big variance occur at same ambient room temp
For example if your ambient is 20c(68f) at you normally run at 25c and
then it raises to 35c at same ambient for no explainable reason
then there is reason to be concerned
basically if your idle is below 40c then dont worry
It is Load that matters since that is where the real accuracy of the Digital Temperature Sensor (DTS) is.
So for testing purposes really only Load Temp matters
Intel has stated that at idle there can be a normal variance of + or - 10c
The great misconception of temperature monitoring is that idle temp
unless it is properly calculated by establishing a baseline to the correct TJ Max
(which Intel has always been pretty vague about TJMax and stated to rely on
TCase instead which there specs use as max)
To have proper measurement the correct TJMax is needed
What it all comes down to is to use all three Tmonitors (HWmonitor,Coretemp and Realtemp) and run a load test (i recommend Prime95 small FFTS)
The reason for using Prime95 is that it is the most popular loader
so there is many examples of idle and load temps for different CPUs
based on using Prime95 to use as a comparison to your own numbers
Basically research you cpu and see what others are getting as temps with systems that are properly running and then use that as a baseline parameter to compare your own readings
Whew glad I am a fast typer