Re: say what?

"But you make it sound like heat is programmed to follow the heatsink from the die."

Well, actually that's exactly what is happening. Copper/Aluminum have way lower thermal resistances than air. Because of this, the heat will flow from the CPU to the heatsink a lot easier than from the CPU to air. Think of it as an electrical circuit with two wires in parallel. One wire has a huge resistance, while the other has a tiny one. In this case the current will almost entirely flow through the low resistance wire, and not the high resistance one. It's the same for heat flow.

Now to answer your question about rising temperatures when the CPU is under a full load. The CPU temperature is a function of heat dissipation and heat production. The final temperature will represent some sort of equilibrium between these two competing factors. As you increase the load on the CPU, it will produce more heat and so it is natural that this equilibrium temperature be higher, since you correctly state that the cooling power of the heatsink is constant (for a set air flow). You could write a some equation like (although this is surely missing many things):

Temperature = Heat generation - Cooling power

I'm sure that you understand all of this though. What you fail to realize is that the CPU surface, just by itself, also has a fixed cooling power. And this cooling power is proportional to the surface area of the system. You might argue that the airflow accross the CPU surface is larger without the heatsink because it doesn't get trapped inside the fins of the heatsink, but even if this was the case (and it isn't) and somehow the airflow was magically increased by a factor of 10, this wouldn't ever be able to compensate for the loss of surface area which would be on the order of 100X - 1000X times!! So, if you increase the load on your CPU, and it doesn't have a heatsink, the temperature will rise even more.

Let me ask you this: if the heatsink is so flawed, why do you think that people even bother using it? If you are so sure that it is a hinderance, why not remove it completely? Don't actually do this though! Now I believe you when you say that your systems are more stable than before, but this has to be due to some other issue. You are misinterpreting your results. Ludicrous had a few good theories. So the question should actually be: why did this work for you? Did you do anything to the heatsink/fan? Perhaps you cleaned them, thus making them more efficient.

Anyway, there's nothing else to say about this. I hope that things are a bit clearer now.

Cheers,

Hugh