Digital computers have reached their peak, the chips cannot be etched smaller than the wavelength of light and they reached that stage years ago, hence the search for more power with multiple cpus on a chip.<quoted text>
The sheer power of such a machine, but yeah, the noise issue is one that has slowed it down considerably as well. Mostly it's that people who code, like me, have become comfortable with binary, it's easy and static so we don't have to worry about precision. Most people don't know enough about computers so they just shrug at the notion not even considering how much computing power such a thing would conceivably possess. If I had a lab for development I'd work on it myself, but alas, you need a bit more funding for hardware tech development than other sciences. I think that analogue computers are the next step, we've reached a ceiling in our digital tech and now we're just adding more processors to the machines instead of actually improving the processors. Oh, and we are improving the PCB layouts, using evolutionary algorithms, that's fun to watch in real time, the computers working out the best PCB layouts is a thing of poetry.
My business (3D animation and graphics) is very computer intensive, the requirement and expectation is now far outstripping the ability of digital processing.
People like Pixar build there own mainframes from hundreds of the most powerful rack PCs simply because there is nothing to do the job otherwise. Itís rather reminiscent of a server room using a custom OS dedicated to one single task.
Perhaps analogue machines will bring about a solution, I hope so because these old SGIís we use are literally getting worn out and there is nothing economically viable to replace them.