Even if your 3GHz 4 cores have a decent amount of cache and can perform their computations without going down the memory bus bottleneck? Remember, the bottleneck would be even worse, because you didn't mention the memory would be twice as fast as well. And, of course, the rest of the buses and peripherials would also be affected, so all waits for memory and for external I/O would, for become effectivly doubly as expensive, as seen by the processors.
Of course, you could say that it'd be nice to have all of the computer's components continue increasing in speed. Well, that'd bring another problem: Motherboard sizes. Because at 6GHz, light speed becomes a limit as well: If, speaking in round numbers, light travels ~300,000,000 meters per second, then it takes 3.33x10^-9 seconds for it to travel one meter. At 6GHz, light travels 50cm per clock cycle. I know I'm comparing apples and oranges here, as electrons don't "move" along the wire, but still — Signals will only travel fractions of that distance on an electronic circuit.
Yes, it could be easier to keep both cores happily going along without programmers having to learn to master concurrency. But we are hitting physical barriers, They do not give way easily.