Home / Technology / AI & IT / Photonics based supercomputers enabled by New breakthroughs

Photonics based supercomputers enabled by New breakthroughs

Moore’s Law which stated that the number of transistors on a chip will double approximately every two years has been the driver of semiconductor industry in boosting the complexity, computational performance and energy efficiency while reducing cost. The increase in computing performance of electronic computers is becoming more and more difficult, as dimensions approach nano meter ranges. The solution has been using parallel processing with multiple cores. While this has enhanced performance, it has led to the problem of rising energy consumption due to enhanced communications required between these cores and outside components. So much so that the communication on and between chips is now responsible for more than half of the total power consumption of the computer.


Researchers are now developing Optical computers. While the computers of today use transistors and semiconductors to control electricity, the Optical Computers of the future may utilize crystals and metamaterials to control light particles called photons. Optical computers promise to be superfast since light travels at 186,000 miles per second. In a billionth of a second, one nanosecond, photons of light travel just a bit less than a foot, not considering resistance in air or of an optical fiber strand or thin film. The Electronic computers are relatively slow, and the faster we make them the more power they consume,the future optical computers shall  also be energy efficient.


Earlier, the realization of optical computers faced many challenges because of the packaging difficulties associated with free-space coupling and holographic interconnects, and the difficulty in shrinking optical devices.


Optical Computers

An optical computer is a computer that performs its computation with photons as opposed to the more traditional electron-based computation.  There are two different types of optical computers, Electro-Optical Hybrid computers and Pure Optical computers.


Information gets sent in from keyboard, mouse, or other external sources and goes to the processor. Electro-Optical Hybrid Use optical fibers and electric parts to read and direct data from the processor. Information is sent using Light pulses instead of voltage packets. Processor then sends the information through logic gates and switches to be programmed. Processors change from binary code to light pulses using lasers. Information is then detected and decoded electronically back into binary.


The information is then sent through different fiber optic cables depending on it’s final location. Some information will be sent to the holographic memory, where it will then be saved. After information is saved and the program would like to use it,the program sends a command to the processor, which then sends a command to receive the information.The program receives the information and sends a signal back to the processor to tell it that the task is complete.


Pure Optical Computers  have no electron based systems or  require conversion from binary to optical thereby greatly increasing the speed.  They use multiple frequencies, and information is sent throughout computer as light waves and packets.

Electronics-Photonics Hybrid computers

“Entirely optical computers are still some time in the future,” says Dr. Frazier, “but electro-optical hybrids have been possible since 1978, when it was learned that photons can respond to electrons through media such as lithium niobate. Newer advances have produced a variety of thin films and optical fibers that make optical interconnections and devices practical.


We are focusing on thin films made of organic molecules, which are more light sensitive than inorganics. Organics can perform functions such as switching, signal processing and frequency doubling using less power than inorganics. Inorganics such as silicon used with organic materials let us use both photons and electrons in current hybrid systems, which will eventually lead to all-optical computer systems.


“What we are accomplishing in the lab today will result in development of super-fast, super-miniaturized, super-lightweight and lower cost optical computing and optical communication devices and systems,” Frazier explained.


Future of Computing: Russian Scientists Create Laser-Based Supercomputer

All-Russian Scientific Research Institute of Experimental Physics in Sarov   have developed a  unique optical computer is believed to have immense advantages over traditional computers, opening up new doors as far as raw computational power and energy use are concerned.


In optical computing, processing is carried out through the interaction of laser pulses, rather than electrical signals, as in conventional computing. The optical computer is divided into electric and photonic sections, with machine code translated into laser pulses. Photons then pass into a photonic processor, where laser pulses interact, allowing for logic operations like the ones in traditional computers to be carried out. After that, laser beams leave the processor and return to the computer’s electric portion, where optical information is converted back into electric-based information, becoming accessible to the user.


The Institute of Experimental Physics’s chief research scientist, Sergei Stepanenko, the project’s creator, told Sputnik that optical computers can solve mathematical problems that are beyond the scope of traditional semiconductor-based ones.
He pointed out, for example, that photonic technology allows for a reduction by tens or even hundreds of thousands of times in the amount of energy necessary to achieve the same performance as a conventional electrical computer.


“Where a supercomputer will require a building the size of a football field, an optical computer can achieve the same performance in a space the size of a half-liter coffee mug and have a heat output of about 100 watts – less than an electric kettle,” Stepanenko explained.


Computer scientists around the world have been working on the concept of photonic computers for many years, but have been unable to come up with tangible results for a variety of reasons, including the nonlinear process in which multiple signals have to interact, the weakness of light waves compared to electromagnetic ones, and other technical issues.


However, the Experimental Physics Institute’s workaround proposes a new scheme for the optical computer’s work, with transitions between the optical and electrical-based components of the computer being performed as rarely as possible so as not to waste time and energy.


The theoretical maximum computational capacity of the photonic computer created by the Institute of Experimental Physics can be up to 50 petaflops (i.e., 50 quadrillion floating operations per second), with this peak power requiring just 100 watts of electricity. For comparison, 100 watts in a traditional computer would allow an output of about 5 teraflops, i.e., 10,000 times less. Furthermore, scientists believe that the power of their optical machine can be even further increased through reductions in the length of the light wave.


Practical applications for optical computers can include everything from mathematical problems to the study of genetic code. Given their comparably low use of power, it wouldn’t be unreasonable to assume that they could be used in environments with limited access to electricity, in remote areas or in space.

Optical memory cell achieves ‘record’ data storage density

Researchers from the universities of Oxford, Exeter and Münster have demonstrated a new technique that can store more optical data in a smaller space than was previously possible on-chip. This technique improves upon the phase-change optical memory cell, which uses light to write and read data, and could offer a faster, more power-efficient form of memory for computers.
The scientists describe their new technique for all-optical data storage in the journal Optica.


Rather than using electrical signals to store data in one of two binary states as with conventional electronics-based computers, the optical memory cell uses light to store information. The researchers demonstrated optical memory with more than 32 states; the equivalent of 5 bits. They say that this development is an important step toward an all-optical computer.


Research team leader Harish Bhaskaran from Oxford University’s Department of Materials commented, “By bringing the speed-of-light-based data transmission to the circuit board, our all-optical memory could enable a hybrid computer chip that interacts with data both optically and electrically.”  The new work is part of a large Horizon 2020 project called Fun-COMP (Functionally-scaled Computing technology), which has brought academic and industrial partners together to develop groundbreaking hardware technologies.

Phase change

The optical memory cell uses light to encode information in a phase change material, a class of materials used to make re-writable CDs and DVDs. A laser heats portions of the material, which causes it to switch between states where all the atoms are ordered or disordered. Because these two states exhibit different optical indices of refraction, the data can be read using light.


Phase change materials can store data for a long time because they remain in the disordered or ordered state until illuminated again with the specific type of laser light originally used to write the data. Mixing different ratios of ordered and disordered states in an area of the material allows information to be stored in a continuum of levels instead of just a zero and a one as in traditional electronic memory.


The researchers accomplished the increased resolution by using a new technique they developed that uses laser light with a single, double-stepped pulse — two pulses put together into a rectangular-shaped pulse — to precisely control the melting and the crystallisation of the material.


Multi-level memory

The researchers showed that they could use their approach to reliably encode data on 34 levels, which is more than the 32 levels necessary for 5-bit programming. “This accomplishment required understanding the interaction between the light and the material perfectly and then sending exactly the right sort of laser pulse necessary to achieve each level,” said Bhaskaran. “We have solved an extraordinarily difficult problem.” The new technique could help overcome one of the bottlenecks limiting the speed of today’s computers: the link between the processor and the memory.


“A lot of work has gone into improving the communication between these two units using fiber optics,” Bhaskaran added. “However, linking these two units optically still requires expensive electro-optical conversions at both ends. Our memory cell could be used in a hybrid optical-electrical setup to eliminate the need for that conversion on the memory side by allowing data to be stored and retrieved optically.”


Next the researchers want to integrate multiple memory cells and individually program them, which would be required to make a working memory chip for a computer. The research groups have been working closely with Oxford University Innovation, the University’s innovation arm, to develop commercial opportunities arising from their research on photonic memory cells. The researchers say that they can already replicate the devices extremely well but will need to develop light signal processing techniques to integrate multiple optical memory cells.


All Optical Processing

All-optical signal processing has been considered a solution to overcome the bandwidth and speed limitations imposed by conventional electronic-based systems. Over the last few years, an impressive range of all-optical signal processors have been proposed, but few of them come with reconfigurability, a feature highly needed for practical signal processing applications. Following the electronic component design strategies, many equivalent photonic signal processors, such as temporal differentiators, temporal integrators, Hilbert transformers and ordinary differential equation (ODE) solvers, have been proposed and demonstrated.


Two very relevant examples of these fundamental devices are temporal differentiators and ODE solvers. Temporal differentiators can be used to perform real-time differentiation of an optical signal in the optical domain and have been applied to ultrafast signal generation and pulse characterizations. ODE solvers play an irreplaceable role in virtually any field of science or engineering, such as automatic control and temperature diffusion processes, writes Ming Li from Institute of Semiconductors, Chinese Academy of Sciences, No. 35, Tsinghua East Road, Beijing, 100083, China, and others.


Andrea Blanco-Redondo and Dr Chad Husko from CUDOS (ARC Centre of Excellence for Ultrahigh bandwidth Devices for Optical Systems) at the University of Sydney’s School of Physics have observed an on-chip soliton compression in a silicon photonic crystal for the first time. Solitons are nonlinear waves that propagate through a medium undistorted and their discovery shall enable development of nonlinear devices in silicon and all-optical processing systems.

UK-based startup Optalysys promises optical processors for supercomputers

Optalysys has completed 320 gigaFLOP optical computer prototype, targets 9 petaFLOP product in 2017 and 17 exaFLOPS machine by 2020 .


“Optalysys’ technology applies the principles of diffractive and Fourier optics to calculate the same processor intensive mathematical functions used in CFD (Computational Fluid Dynamics) and pattern recognition,” Dr. Nick New, CEO and founder of the company explained. “Using low power lasers and high resolution liquid crystal micro-displays, calculations are performed in parallel at the speed of light,” New further described.


The Optalysys Optical Solver Supercomputer reduces energy footprint so it can be considered as eco-efficient as well. No need for special power as it only needs a standard mains power supply.


As for the running cost, the super processor only costs £2,100 every year. This is way cheaper than the fast supercomputer available in China, the Tianhe-2, developed by the National University of Defense Technology. The computer alone costs $320 million but with a $21m annual running cost.


Luminous Computing raises $105M to build a photonics-powered AI supercomputer

Startup Luminous Computing Inc., which is developing a new kind of supercomputer for running artificial intelligence models,  said in March 2022 that it has closed a $105 million funding round.


The vision behind silicon photonics is to leverage the fact that light travels faster in order to perform calculations more quickly than is possible with a traditional processor in which data is encoded as electricity. That’s the vision Luminous is working to implement.


Luminous intends to use its silicon photonics technology to build a supercomputer specifically optimized for AI workloads. Using proprietary silicon photonics technology to eliminate data movement bottlenecks at every scale, Luminous is completely re-imagining how AI computers are built, resulting not only in order-of-magnitude improvements in performance, but also in drastic simplifications to the programming model.


“Most people who build hardware assume that in order to improve performance, you have to trade off against programmability and cost-efficiency, or just go to a higher-density silicon node,” said Luminous co-founder and Chief Executive Officer Marcus Gomez. “By introducing silicon photonics technology at the heart of computer architecture, we’re not only able to drastically improve performance and scalability, but we’re also able to make it much easier to build huge AI models.”


According to VentureBeat, Luminous has already produced multiple working prototypes of its silicon photonics chip. The startup aims to start shipping development kits to early customers within two years. According to Luminous, its initial target market consists of hyperscale data center operators such as the major cloud providers.




References and Resources also include:





About Rajesh Uppal

Check Also

Advancing Warfare: Joint All-Domain Command and Control (JADC2) for Integrated and Networked Multidomain Capability

Introduction The future of warfare is rapidly evolving, shaped by advancements in technology and the …

error: Content is protected !!