>>31
Intel made processors that melted people's balls and backed off from those blistering (HEH) speeds. Computers are no longer commodity devices if they're hotter than the sun and require heavy equipment to cool.
Since you brought it up: The effective processing speed did
not hit a wall. Intel's >3GHz offerings from around 2004 used a lot of stupid tricks to make it seem like they were faster than they actually were, because as long as they could get away with a higher MHz label, they could sell newer processors. The per-instruction latency in the PIV and friends was actually harmful to overall performance compared to, say, a PIII-based design with the same clock speed.
If they'd done it "the right way" the whole time (ie. relying only on smaller processes for faster speeds instead of doing parlor tricks with the instruction pipeline) you'd see a much gentler slowdown over time.