Important Gaming News

[Hylix Ulyx]

Aha!!! You took the words right out of my mouth....

 

 

....If I worked with optics....

 

 

 

 

Hylix.

[sparhawk]

I imagine an optics-based computer would be horrendously expensive for the reason you mention - quality of light.

 

That's not really an issue. The first computers alos were as big as a house and just as expansive (or worse) with processing power less then a TI-34 now has. Nothing any normal Joe Average could afford. Today this very Joe Average has a PC on his desktop that is so cheap that you can even have it and only use it once a month and it has more processing power then he sensibly could use (apart from games of course).

[Gildoran]

I was under the impression that the main limiting factor of current computers isn't the CPU, but rather the bus and memory speeds. I've heard that in the time it takes to retrieve a few bytes of memory, a modern CPU can execute 60 to 100 instructions. Getting a faster CPU doesn't improve performance much because the speed at which memory is accessed remains the same. Of course, hardware manufacturers try to mitigate that by caching chunks of memory on the CPU and in a small amount of expensive high-speed memory. But the reason for the diminishing returns on better CPUs is because the rest of the computer can't supply the processor with data quickly enough to make use of its full potential. I look forward to optical CPUs, but I don't see them as providing much of an advantage over silicon processors unless bus and memory speeds improve. (is optical memory feasible?)

[Ishtvan]

There's work on optical memory and buffers, using everything from delay lines to bistability in lasers to defect states in photonic crystals. However, I still think all optical processing for PC's is way way in the future, unless some major breakthrough happens.

 

I'm not an expert (just a grad student), but if I had to guess, I'd say we'll first see more optoelectronic components, like optical interconnects between chips (instead of a bulky bunch of wires that heats up).

 

There's just not that much funding being devoted to all-optical computing research right now. Most investors see it as a pipe dream. The military invests some for faster target acquisition and identification, and telecomm invests some. In telecomm, it is more feasible to try out all-optical devices than in the home PC market. For example, it might be feasible to install an all-optical router in the telecomm network that is bulky and uses a lot of power but works much faster to increase overall system bandwidth. Still, there are a lot of sunk costs in telecomm like routers already installed and fibers already in the ground, so a lot of stuff just sort've gets added to a proposed "next generation network" plan that will be built at some unspecified time in the future.