This is extremely unlikely, because for it to be the case you have to have somehow managed to run a CPU with a 20 W TDP so hard that it got that hot.

The sensor isn't being read correctly, and the package temperature is actually much higher.

I know 100% for certain that Linux reports the temperature correctly on that model of CPU because the sensor interface the CPU provides has been around since the AMD K10 days and is very well supported by Linux). Everything else reporting similar temperatures does not rule this out though, because the drivers being used to make the reading may be bad (you can check that case by booting into a live Linux environment and seeing what it says the temperatures are.

This is the best possible case, and it's easy to check (try a handful of other tools for reading these sensors.

The sensor isn't being read correctly and the core temperature is actually much lower.In particular, the possibilities that come to mind are: However, I'm inclined to believe something is wrong with your system due to that insane discrepancy between reported package and core temperatures. Provided you have a working (and properly sized) cooling system and are not somewhere with unusually high temperatures to begin with (40 ☌ or higher), you should not be seeing temperatures that high no matter how hard you push the CPU. Most consumer electronics (including CPUs) are not designed to operate above 85 ☌ for any extended period of time, and most will actually shut down when they get over about 100-105 ☌. You're well into the range that can cause permanent damage to the CPU.

Cpuz temperature cpu