The Visual Cortex is an example of a very very powerful specialized computer. The best way to imagine how powerful it is, is by looking at graphics of simpler visual processors.
If you're familiar with how insects, dogs, and animals see then you have an idea. Some simpler visual processors can sometimes distinguish perspective, sometimes color, distance, patterns, texture, or has a powerful ability to adapt to diverse lighting conditions (photographers should appreciate this the most). You can make a comparison to graphics in computers to those of other visual processors and still come up with some parallels in complexity.
Now here's another kicker, object recognition and context recognition. One thing is to visually construct the world around you (your eyes really do a lot more of inference than you think; if you've been in G+ the visual illusions should point out a lot of its weaknesses and amazing compensation abilities), its another to be able to, without consciously thinking, make decisions based on all the other specialized processing centers: Spatial, kinesthetic, touch, equilibrium etc...
Imagine just driving and the amount of processing going on, if you consider the ability to navigate and drive. Then consider the meta thinking when you strategize and plot a course, then consider that your going a route, using route memory, and multi-tasking (like I always do) like doing problem solving task, developing a composition, or mulling over the complexities of a social interaction.
The graphics alone requires a computer 100x better than a desktop (Complexity 5; passmarks 100,000-999,999) (roughly the computing power to render an high quality image in less than a second continuously at a rate of about 40fps), every additional task that coordinates with the specialized processing requires a processor as sophisticated: part of the processes is coordinating with the visual, part of the processes in the specialized intelligence tasks, and another part coordinating with the other Intelligence. If you believe or adhere to multiple-intelligence theory you have: kinesthetic, spatial, memory, visual, logical, mathematical, linguistic, musical, intrapersonal, and interpersonal. A human brain might be around Complexity 6 (IQ10-11 in GURPS), but it has the capacity to become x10 (IQ12-13) through more effective through mnemonic techniques, conditioning, and education.
The funny thing about orders of magnitude of 10, is that you might have 10-100x better in one intelligence but as you coordinate and implement with all others: communicating ideas, working in a socially complex structure, converting the ideas to multiple mediums, kinda forces some kind of averaging out.
So in around TL10+, a C5 tiny computer is 0.04lbs and a C6 small computer is about 0.4lbs. You can probably fit a small computer in your head and probably about several tiny computers, a neural interface, radio communication.
So the difference in genetics, I can make up for my intelligence by the amount of additional computer system I'll implant. From having one crawl down my spine, another tiny computer, to an organic and regenerate-able flexible computer embedded in one's skin (several lbs of a circuitry), to evasive surgery min-maxing the space inside one's body (as well as finding a way to generate enough power for all of that; like a skin suite that powers through solar, heat and kinesthetics through hydro-distribution).
anyways this is how I try to put computing power into perspective.