Increased size of the inferior parietal lobe, statistically significant increase in the level of glial cells, enlarged Sylvian fissure, vacant parietal operculum. That's what we discovered in Einstein's brain. Unfortunately, Einstein is only one statistical point. Without more genius brains, it's difficult to say what exactly caused his genius. Therefore, let's study -all- the brains of dead people. As many as we can get. The one thing we can say from that is that there are clearly physical differences in brains, although if you dismiss everything different as a defect and irrelevant, I guess that doesn't help much.
Also, as far as the computer comparison goes, I think that's particularly apt. Especially if you consider the practice of binning. Not all CPUs are created equal, even in manufacturing processes as close to identical as we can manage. Some work better. The same thing used to be done with guns. Before leaving the factory, sharpshooters would put a number of rounds through them, and some rifles would be more accurate than others. Those that went past a certain threshold would be put aside and sold separately, as the Winchester "One of One Thousands." It happens with screws, there are various grades to define how close to spec the threads are manufactured.
With the best precision manufacturing capability history has ever known, we still can't manufacture a screw to perfection every time. And you think that there won't be variations among brains?