Optical Processing Pioneer wins Project with DARPA

Cambridge University spin-out Optalysys has been awarded a $350k grant for a 13-month project from the US Defense Advanced Research Projects Agency (DARPA). The project will see the company advance their research in developing and applying their optical co-processing technology to solving complex mathematical equations. These equations are relevant to large-scale scientific and engineering simulations such as weather prediction and aerodynamics.

The Optalysys technology is extremely energy efficient, using light rather than electricity to perform intensive mathematical calculations. The company aims to provide existing computer systems with massively boosted processing capabilities, with the aim to eventually reach exaFLOP rates (a billion billion calculations per second). The technology operates at a fraction of the energy cost of conventional high-performance computers (HPCs) and has the potential to operate at orders of magnitude faster.

In April 2015 Optalysys announced that they had successfully built a scaleable, lens-less optical processing prototype that can perform mathematical functions. Codenamed Project GALELEO, the device demonstrates that second order derivatives and correlation pattern matching can be performed optically in a scaleable design.

Project EQUATE: Light speed mathematical processing
Objective of project EQUATE is to investigate applying the Optalysys optical processing technology towards solving complex mathematical equations that form the basis of large scale simulations such as those used in the dynamical core of earth systems modelling.

A feasibility report describing the outputs of an Optalysys research system performing example mathematical functions will be produced and a roadmap for scaling the technology into a commercial product.

In 2015, Optalysys completed a 320 gigaFLOP optical computer prototype. They were targeting a 9 petaFLOP product in 2017 and 17 exaFLOPS machine by 2020.

The project’s goal is to lay the groundwork for producing optical processing systems that are capable of high-end tasks used in computational fluid dynamic simulation models such as Direct Numerical Simulation[1] and Large Eddy Simulation[2].

The Optalysys technology is currently developed to Technology Readiness Level 5 (TRL5), and is being developed for use primarily in the field of genetic sequencing, as part of a UK government funded project in collaboration with The Genome Analysis Centre (TGAC) in Norwich, UK.

Dr Nick New, CEO and Founder of Optalysys said:

“We are reaching the limits of what traditional silicon-based processors can deliver. Moore’s Law is breaking down and traditional computing methods are approaching their limits in terms of cost and capability. The Optalysys technology is built on the well established principals of Fourier and Diffractive optics but we use them in combination with advanced high resolution microdisplays. We are creating a cost-effective solution that can be scaled beyond the levels of traditional electrical computers and can be integrated with existing desktop and HPC architectures.

We are developing the technology specifically to help speed up research and analysis for organisations that are trying to help solve some of the world’s biggest problems so having the chance to work with DARPA in this way is an exciting step forward for us.”

[1] Direct Numerical Simulation (DNS) is a method used in Computational Fluid Dynamics (CFD), the process of modelling real-world fluid in a computer (weather forecasting, aerodynamics, flame propagation, nuclear chain reactions etc.) DNS is the most accurate of the CFD methods and is also the most computationally intensive, significantly constrained by existing electronic processing limitations.

[2] Large Eddy Simulation (LES) is the simulation of turbulent flows by numerically solving the Navier–Stokes equations. Direct Numerical Simulation (DNS) is more accurate but is computationally expensive and currently prohibitive for practical problems. The main idea behind LES is to reduce this computational cost by making compromises.

What is the basis for the Optalysys technology?

Using diffraction and Fourier Optics, coupled with their novel designs, they are able to combine matrix multiplication and Optical Fourier transforms into more complex mathematical processes, such as derivative operations. In place of lenses, they also use liquid crystal patterns to focus the light as it travels through the system. This means the tight alignment tolerances that exist through the system are achieved through the dynamic addressing in the software.

Optalysys technology is quick at Computational Fluid Dynamics (CFD) calculations, so what?

Optical techniques can make all stages of the CFD process more effective. They are far from a pure optical CFD solver but it can help us generate, and learn from data better. They are always thinking about the full CFD process and learning. This leads to big data, again another exciting opportunity.

An optical CFD solver would be the ultimate project to get involved with and they would love to partner up to do this. Their focus is to prove initially that optical technologies can improve and complement what is already being done digitally.