Yeah, normal, thats EIST at work. CPU gets used for a split second so they jump up instantly, and then back down.
Read this also.
Originally Posted by RealTemp Creator/Author (unclewebb)
Intel designed these temperature sensors to control thermal throttling and thermal shut down and for those purposes, they tend to work excellent. They were never designed to be used to report accurate core temperatures.
The first problem is that TJMax is not a clearly defined value. Intel released a lot of information last year but they now refer to this data as TJ Target. Actual TJMax for any processor can either be equal to TJ Target or in some cases it might be slightly higher than the Intel listed TJ Target. Unfortunately, Intel did not provide the user community with any information about how much TJMax tends to vary. Based on my testing and reports from users, a 5C difference doesn't seem too unusual with 65nm and TJMax might vary by 10C in the 45 nm CPUs. The result is that two CPUs with the exact same model number might have different TJMax values and it also seems that there are situations where cores on the same CPU can have different TJMax values. For the 45nm processors, TJ Target and TJ Max seem to be similar but for the older 65nm CPUs, actual TJ Max can be 10C or higher for many processors.
Starting with Core i7, Intel has improved this situation by including the TJ Target goal within each CPU. This information is still only TJ Target though so there is no guarantee that the actual TJMax of a core is exactly equal to the value stored within each core of these processors but at least this is a step in the right direction.
The next issue that effects all of these sensors is slope error. That is when the data coming from these sensors moves at a different rate than the core temperature is changing at. The Calibration feature in RealTemp is used to compensate for sensors that move along different temperature curves and once again the Core i7 sensors seem to be greatly improved with only very minimal slope error in the normal temperature range.
The biggest problem is that the sensors used on the 45nm Core 2 Dual and Quad core processors can become saturated or can become stuck at lower temperatures. As your core temperature decreases, they might reach a temperature where they continue to report the exact same temperature even though your CPU continues to cool down. With a stuck sensor, there is nothing you can do about it and it prevents you from doing an accurate calibration of your temperature sensors.
With all of these problems, it's obvious why Intel has never recommended using these sensors to report accurate core temperatures. My opinion is that as long as your sensors are not stuck and you are able to calibrate them then you should be able to get some reasonably accurate core temperatures out of them. The sensors simply aren't designed or documented well enough to report 100% accurate core temperatures. All software is going to be limited by these problems. RealTemp is the only program I know of that is fully adjustable and tries to make sense out of the sometimes random looking and chaotic data coming from Intel's digital thermal sensors.