As in communications, optical computing promises higher speeds and lower power consumption. Although optics has become critical for modern datacoms, optical computing has never seriously taken off, mainly due to a lack of commercially viable optical transistors. Classical optical computing remains an area of research, and some forms of all-optical signal processing, which verge on optical computing have been realized using integrated photonic circuitry. In high-speed applications that warrant dedicated hardware such as synthetic aperture radar, or auto-correlation circuits this is a trend to watch. Examples like these demonstrate the cutting edge of application-specific photonics which is well supported by Lumerical via foundry PDKs in INTERCONNECT and Electro-photonic design automation EPDA workflow supported through Cadence Virtuoso.
More recently the discussion of optical computing has pivoted towards photonic quantum computers which hold the promise of being CMOS-compatible. Such a fabrication heritage would reduce cost and increase device density allowing the number of qubits to rapidly scale. Furthermore, some proposed variants would be able to operate at room temperature or at least above milli-Kelvin which would make them more robust and versatile than current commercially available superconducting designs. Although, FDTD is a classical solver, in quantum optics the classical fields typically serve as the input values to be quantized and so accurate classical EM analysis is still crucial.
For those interested in classical and quantum photonic computing many standard passive, and active devices as well as circuit components will be crucial. Challenges in classical optical computinginclude developing transistors and practical memory buffers. Transistors are nonlinear electrical devices which serve as the basis of all modern logic gates. Various phenomena that utilize optical nonlinearity have been investigated to replicate this behaviour optically. FDTD provides a powerful and general technique for investigating such mechanisms. Developing reasonable optical memory buffers beyond a delay line using some ideas discussed in Data Storage would allow photonics to move beyond synchronous logic.