[airspirit]

I like the instant calculus in my head thingy. I think that is where much of the confusion is coming from here. The more efficient the block is, the more heat that will be transferred to the water rather than dumped through other means (air, board, etc). That is the meaning of that graph. If you dump a larger percentage of the identical heat load into the water (see above), the water will be warmer. Though the radiator will work more efficiently because of the larger dT, it will NOT cool the water to below the temp of the lower heat input. Therefore, you will have warmer water. Not a big difference, but a difference nonetheless.

I don't think any of what I said qualifies as a "tilt". Theory is only as good as what it's applied to. As far as heat travelling from higher to lower regions, that still applies. The die is hotter than the water, which is then hotter than the rad. Everything fits. A more efficient block will assist the heat in taking the path of least resistance: the path to the water. This is just like reducing the restriction on a parallel circuit: it will cause more energy to flow through that particular branch. Since more of that thermal energy will be sent to the water as opposed to other distribution sources (passive cooling), more energy will go to the water. This phenomenon can not be disputed as it holds perfectly to all the theory out there. Plus, it doesn't magically make heat disappear, unlike some of the other ideas being thrown about.

I respect your expertise, guys, but I think you're wrong this time.