Hmm, that is interesting, however not nearly as different as it should be. On CPUs with 4 different sensors, it is extremely rare to have all 4 cores the exact same temp. I have seen them different, even on my PII 940. However, it is never more than one read cycle. I wonder if it may be due to the program refreshing (taking a new sample) at the same time the sensor is updating. So rather than recording temperatures in different places, it is reporting temperatures at different times.
That sounds reasonable. My Athlon 64 x2 5600+ cores rarely varied, but on my brothers Athlon 64 x2 5050e I noticed about a 5C degree difference between the cores on load.