Optic computing

[PatricBernard]

I have recently become extremely interested in the concepts of optic computing, the process of eliminating electrical components from the computing functions in a computer, mainly in the CPU.

Throughout the day I have been thinking of possibilities that could make this vision a reality, but their are just a few things I cannot even imagine how to solve. For instance, the memory functions of a computer, or how logic gates would work.

The final thoughts I had during they say were that although this would compute information given from a user, to data, the same way an every day computer would, this would not be the same thing. The whole computation process would have to be re designed and devoped from scratch.

Anyone have any input on this? Just tossing ideas out there, getting everybody's creative minds working!

 

[MacGyverS2000]

Plenty of optical memories out there (CDs/DVDs are optical) or being worked on (holographic memory) in the academia side of things. I see no reason to redesign the overall architecture just because some things more from electrical to optical.

Is there something specific you were after?

Dan - Owner

http://www.Hi-TecDesigns.com

 

[IRstuff]

Holographic memories, unfortunately, are looking the way of gallium arsenide technology, i.e., "GaAs is the technology of the future... and always will be." GaAs was supposed to supplant silicon integrated circuits in the mid 90s; there were companies that started up to service the production needs of that technology. Nothing much ever materialized, except for niche markets. My previous company dumped millions into it, developed custom GaAs processors, etc., to little ultimate benefit.

Holographic memories were "being worked on" 20 yrs ago. Lots of fancy claims, but the precision of the readout technology is so tight that the costs have yet to approach any comparable competing technology. The sheer density is impressive, but that stress bottlenecks elsewhere in the chain that have yet to be conquered.

Optical computation techniques were described nearly 30 yrs ago, to no avail. There haven't been any problems that can justify the expenditures for development. Moreover, other approaches may prove to be easy to crack, like quantum computing.

TTFN
faq731-376

 

[PatricBernard]

The only things I was after was overcoming Moore's Law, and also creating a better way for computers to operate.

If you think about it, Right now, all the information in a computer has to go through a single point, (Well multiple with multiple cores and whatnot) the CPU. However, This is like cramming an 8 lane highway into one lane, and then opening it back up to 8 lanes again. If you look at how the human brain works (Which is also a very advanced computer) All the information goes to exactly where it needs to go, like a spider web of information highways. That would be my main focus for the architectural redesign. To make computers think more logically than they already do.

My reason for choosing light as a source to transfer info would be that it is the fastest form of communication available to us.

IRStuff, I do see your point in how quantum computing could be easier to approach and crack (And I think they have already made huge leaps and bounds in that direction). My only issue with this, is wouldn't quantum computing also have to deal with Moore's law as well?

 

[MacGyverS2000]

If you want the fastest form available, quantum entanglement is what you're after... light is constrained by its speed (even in a vacuum). Entangled bits don't need to travel anywhere.

Dan - Owner

http://www.Hi-TecDesigns.com