Degrade all known with time.
The thermal conductivity they vary - if you take the net conductivity, the difference can be an order of magnitude.
If you take the real efficiency difference between the best and worst samples does not exceed 10-15%.
In fact the KPT-8 provides a sufficient level and copes with their work in most cases. To take something more thermally conductive is only if you are overclocking, and want to achieve some unique results.
If you paste is rapidly degraded, then often the problem is not the quality of thermal paste and mount the heatsink. Not provided by normal and uniform clamping of the heatsink to the processor.
luella.Kie answered on June 7th 19 at 14:32
From all that tried it (complete with different radiators, KPT-8, Alsil), cooler was Arctic Cooling MX-4. In comparison with KPT-8 the temperature is below 5-10 degrees (especially noticeable on hot cards). Long does not harden. But we must understand that it needs thermal paste (any) applied correctly - a very thin layer, evenly, with no gaps. I use a plastic card. And the radiator should be well pressed against the cooling surface, and should be excluded the possibility of bias. Well, the radiator should allow high-quality heat dissipation. When hot the CPU and the little heatsink and thermal grease of any kind will not save.
For reference, the coefficients of thermal conductivity W/(m·K):
KPT-8 - 0.7 - 0.65
Alsil-3 - 2
Arctic Cooling MX-4 - 8.5.
The lack of Arctic Cooling MX-4 is still there. It is the price.
Nico_Nolan answered on June 7th 19 at 14:34
For laptops use amperin ss100 is a great thermal paste, cheap. Tried KPT-8, worked worse than the old one dried up.