cnhoff et al,.....(just correcting the physics)
The watts/heat created always comes out...the issue is what does the internal temp have to rise to in order to force the watts out, ie, the equilibrium condition.
Water flow analogy...visualize water flowing at a constant flow rate into a tall "vented" standpipe that has a partially open drain valve. The level will rise, increasing pressure and drain flow until drain flow = input flow.
Now close the drain valve a little bit, increasing resistance. The drain flow decreases temporarily. So the level will rise, gradually increasing pressure some more until drain flow = input flow again.
The change in level/pressure is analagous to the change in cpu temp if you changed to a poorer, higher thermal resistance heatsink.
You don't change a cpu's power dissipation when you change the hs, you change the resistance to heat flow. So temp changes because of that resistance change at constant heat flow, just like the constant water flow/vented standpipe/drain valve analogy.
John C.