Optical Computing: Exploring the Possibilities

In summary: It also has the potential for faster CPU speeds, but currently faces constraints such as longer wavelengths and higher energy. However, some believe it has the potential to surpass quantum computing in terms of practicality and application.
  • #1
Claude Bile
Science Advisor
1,479
26
People seem very quick to proclaim Quantum Computing as the solution to faster computers, however the development of Optical Computers rarely rate a mention. The motivation behind this thread is to give some awareness as to what optical computing is and what it is capable of and hopefully invite some discussion.

The first widely used photonic waveguides were optic fibres which were developed around 1980. Optic fibres are an excellent means of data transfer over long distances. Optic fibres have also been used to construct some very clever photonic devices (splitters, multiplexers and circulators for example), however there is a fundamental limitation in using optic fibres in minaturised circuits.

This limitation is due to the way optic fibres guide light. Fibres guide light through total internal reflection (devices such as these are often termed 'index guided'). If one bends a fibre too much (as is necessary when minaturising photonic circuitry), the fibre will no longer guide light efficiently.

A solution to this problem emerged in the form of photonic crystals, that guide light using the phenomenon of photonic bandgaps (which are not unlike electronic bandgaps). If light of a particular frequency that falls within a photonic band gap, is incident on a photonic crystal, the photon will be reflected (the photon effectively satisfies Bragg's condition). If one creates a defect through the crystal, the photon will be guided. This method of guiding light does not suffer from the same drawbacks as index guided devices, however there is the issue of fabricating such crystals.

In order for a band gap to be present at a particular wavelength, the dielectric constant of the crystal must be modulated in dimensions of the
order of half that wavelength. Thus to guide wavelengths commonly used in optical devices (1.55 microns), the modulations must be of the order of hundreds of nanometers.

In addition to guiding light on a small scale, there needs to be some sort of way to store data optically (i.e. the optical equivalent of a transistor). One such proposed method is holographic storage, however it suffers from the drawbacks of being short-lived (about a day or so) and must be kept very cool (around 100 K) to work properly. Nonlinear Fabry-Perot cavities are another proposed solution, however they are expensive to make and difficult to align and operate. The difficulty is in that most optically bistable materials are only bistable under very specific conditions and usually at a low temperature.

Despite these experimental obstacles, optical computing, in my opinion, is closer to being a reality than quantum computing.

Okay, I'm running out of time, so I will wrap up discussion here for now. Please feel free to discuss anything mentioned in this post, or if you have an opinion regarding optical vs quantum computing, I would be interested to know.

Regards,
Claude.
 
Computer science news on Phys.org
  • #2
Could you please describe in a few simple words what is quantum computing?
 
  • #3
Hi Claude Bile, your presentation of optical computing seems interesting and promising. Since light travels at the maximum speed, it would be able to transfer data at tremendously high speeds. How could this technique be used for computer microprocessors? Till now microprocessors have reached a maximum clock rate of 3GHz, all this achieved with electricity. Research to speed up microprocessors has focused on shortening the paths that electricity follows in a fetch-execute cycle. This is a complex task, since there is a cobweb of paths and circuits and it is difficult to predict the route. Other techniques have been to use multiple microprocessors working in parallel, piping, using fast cache memories and so on. These trends however promise only minor advances. I don't know why so little research is being made in the so-called optical computing, which might result in tremendous CPU speeds, maybe several THz or further.
Apart from optical busses (paths) we would also need optical memory. Your suggestion of holographic memories might be a solution, but isn't it slow? It might prove to be a major bottleneck. What about transistors, which also need to be optical, to implement switches for logic gates (AND, OR, NOT )? I don't know any. Also all the computing system has to transit from electric to optical (including peripherals). If this could be done we would have achieved tremendous results in computing speeds.
In any case I suspect that in practice we wouldn't need much faster microprocessors. Major bottlenecks have become memories like RAM, ROM, etc. Also statistics show that with modern computers and advances in CPU speeds most of the time is spent in performing Input/Output operations in peripheral devices, especially the hard disk. These peripherals have become a major bottleneck. If we need to speed up the flow of data (like in CPUs) we also have to improve in parallel the performance of memory, disks, printers, network connections (where we already have introduced optical fibers), and so on. What's your opinion?
 
  • #4
This is definitely an interesting and thought provoking subject. My opinion regarding optical computing is that most people, including myself, are skeptical due to two major reasons: compared to electron, the wavelength of a photon is too long, and also its energy is too high. Both place a constraint on how small the end device can be, and size do matter, because bulky optical devices consume too much power. In addition, all of the optical devices demonstrated to date (or those I'd read about) that are capable of performing logic operation are based on inteferometric configuration, which again places a lower limit on the device size.
Apart from that, popular research topic in quantum computing, like spintronic for instance, interface with current microelectronic technologies much easier compared to optical computing. So I think as far as computing is concerned, electron still has the upperhand. Photons work best in transmitting information, but my gut feeling is that being boson, it's much more difficult to manipulate compared to electron.
 
  • #5
OldTee said:
This is definitely an interesting and thought provoking subject. My opinion regarding optical computing is that most people, including myself, are skeptical due to two major reasons: compared to electron, the wavelength of a photon is too long, and also its energy is too high. Both place a constraint on how small the end device can be, and size do matter, because bulky optical devices consume too much power. In addition, all of the optical devices demonstrated to date (or those I'd read about) that are capable of performing logic operation are based on inteferometric configuration, which again places a lower limit on the device size.
Apart from that, popular research topic in quantum computing, like spintronic for instance, interface with current microelectronic technologies much easier compared to optical computing. So I think as far as computing is concerned, electron still has the upperhand. Photons work best in transmitting information, but my gut feeling is that being boson, it's much more difficult to manipulate compared to electron.
OldTree, I cannot follow your reasoning, or you are unclear. You say that Both place a constraint on how small the end device can be, and size do matter, because bulky optical devices consume too much power.. I don't know why this could be. The end devices don't take the power from these light signals, but from other sources. These light signals serve only for data and can be amplified to any magnitude we need.
 
  • #6
Opps..apologies for being unclear. By end devices I'm actually referring to the final design of these optical signal processing chips. Just take amplifier as an example, semiconductor optical amplifier is much much larger than an electrical op-amp. And the power it takes to amplify optical signal is much higher than what it takes to amplify electrical signal for a given magnitude.
 
  • #7
OldTee said:
Opps..apologies for being unclear. By end devices I'm actually referring to the final design of these optical signal processing chips. Just take amplifier as an example, semiconductor optical amplifier is much much larger than an electrical op-amp. And the power it takes to amplify optical signal is much higher than what it takes to amplify electrical signal for a given magnitude.
Maybe, I didn't know that. But does it have any link with electron wavelength and energy being smaller than that of the photon?
 

Related to Optical Computing: Exploring the Possibilities

1. What is optical computing?

Optical computing is a form of computing that uses light instead of electricity to perform calculations. It utilizes the properties of light, such as its speed and ability to be manipulated, to process and transmit data.

2. How does optical computing differ from traditional computing?

Traditional computing uses electrical signals to process and transmit data, while optical computing uses light. This allows for faster processing and transmission speeds, as well as lower energy consumption. Optical computing also has the potential for higher computational power and parallel processing.

3. What are the potential applications of optical computing?

Optical computing has the potential to be used in a variety of applications, such as data processing, communication networks, and artificial intelligence. It could also be used in high-performance computing, image and video processing, and cryptography.

4. What are some challenges facing optical computing?

One of the main challenges facing optical computing is the development of reliable and efficient components, such as optical switches and detectors. Another challenge is integrating optical components with traditional electronic systems. Additionally, there is a lack of established standards and protocols for optical computing.

5. What advancements have been made in optical computing?

There have been significant advancements in optical computing, particularly in the development of optical components and materials. Researchers have also made progress in integrating optical and electronic systems, as well as in developing algorithms and protocols for optical computing. However, further research and development is still needed to fully realize the potential of optical computing.

Similar threads

  • Materials and Chemical Engineering
Replies
1
Views
944
Replies
1
Views
1K
Replies
7
Views
2K
Replies
1
Views
1K
  • Introductory Physics Homework Help
Replies
15
Views
1K
Replies
2
Views
806
Replies
8
Views
2K
Replies
8
Views
3K
Replies
18
Views
1K
Back
Top