Quote:
Originally Posted by Robo
I think Kraetos and chucker are right in that the 2013 Mac Pro design required the thermal load to be evenly balanced among a CPU and two GPUs in a way that just wasn't where the market ended up going. I don't think it was just a "they should have just spun the fans faster" thing; they made a fundamental misjudgment on where the future was going for high-performance computing. They bet on (and spent a ton of money to build a bespoke saddle for) the wrong horse.
|
Phil Schiller is literally making shit up. Heat radiates REALLY quickly from the cores to the heatsink fins. Air ablates the heat from the fins and out of the enclosure. This is true regardless of heatsink design. If the heatsink is one piece, all the better. That distributes idle and low loads and allows the fan to stay at lower RPM longer. Hell, if thermal designs were dependent on faster radiation, then methods like liquid cooling wouldn't even be feasible.
Besides like I already said, pin compatible solutions with lower TDPs existed. Haswell E was more efficient than Ivy Bridge E, FirePro S9150 more efficient than FirePro D700. Apple chose not to use them.