From a theoretical point of view: NO change in temperature.
Yeah a slight change in temp for a non-ideal gas, but it's mostly entropy.
Increasing the convection in the heat pipe isn't a huge increase in efficiency so long as these are properly mounted.
Are you really sure about that? Hmmm... how does one calculate the mean free path for gas particles again? [
Yes - I an extremely sure about that. The mean-free-path(MFP) has little to do with gas conductivity. Instead the only factors that you can realistically tweak are to choose gas molecules with less mass, smaller radii and operate the gas at a higher temp (which defeats the whole purpose).
Now I know someone will note that gas thermal conductivity is proportional to (n * <v> * MFP), where n is the molecular density and v the mean velocity, *BUT* if you factor out the 'n' and the 'MFP' you just get a relation where conductivity varies in relation to sqrt(T) [square root of the absolute temp]. This fact rather surprised me when I was desigining some equipment for very high altitude (65k feet+) use 2 years ago.
I see that modern heat pipes use a "wick" or scintered material to carry the fluid phase, while older models used gravity to drive the liquid to the hot end. Either way - the improvment in heat-pipe conductivity does not help in cooling a CPU.
I have read that cheap heat pipes have ~80 times the conductivity of a similar copper volume, and very well built ones perhaps 1000 times. This really doesn't help us. It's like saying the 0.5 watts lost in a circuit breaker could be halved if we use two in parallel - it's true but it is not a useful gain.
Heat pipes can help us transport the heat far more (say 80 times more) effectively than copper, which is already pretty good. We ultimately must transfer the heat to the environment, primarily air and that's the real problem here. (more to follow).