AMD's don't currently have an on-die thermistor, so any temp measurements that will be given have to be qualified by many factors. First off its measured by a thermistor that is measuring the backside of the processor, which is ceramic, which makes all your temps subject to error. Each motherboard varies in both the thermistor placement, but also on the temp. correction used in their bios to compensate for in-direct measurements.
With all that said, with my 1ghz @ 1.233ghz with an Alpha Pal6035/delta combo averaged 45-50C full load with my thermistor bent up to touch the bottom of chip directly. Mind you those have been in 68 fahrenheit room, and it is now going to go up as summer comes. (I since switched to watercooling and have my same processor @ 1.370ghz @ average 10C less measured by same thermistor.)
Your temps "might" be about the same with the same heatsink, mobo and average case fans. So I kinda test other ways like is the base of my heatsink cool, warm or very warm. If hot you are probably hurting your processor already turn it off. If very warm you might consider getting better case cooling. If warm or cool you are ok. This is not an accurate way by any means, but if you use your estimated temperature to stability(test stability with something like Prime95 and test if your system stays stable when you pump the heat of your video card into case as you run it hard by looping 3DMark2001) you can get a decent idea if you are ok with your cooling or not.