The optical chips promise 350 times the performance of the RTX 3080
Photo Credit: Lightenlligence Still in the demo stage, the Photonic Arithmetic Computing Engine (PACE) integrates electronics and photonics into one single chassis, performing lightning-fast operations and thus achieving extreme acceleration for certain AI, Deep Learning, and Machine Learning compute workloads. PACE accelerates some matrix acceleration applications much faster than NVIDIA's RTX 3080 due to the very nature of its processing elements. It is quite easy to understand: latency, which is the time interval between the moment an event is ordered to occur and the moment it actually occurs, is much, much lower in the Lightelligence system. This is the advantage of data traveling at the speed of light.
To achieve this goal, Lightelligence knew it would have to focus not only on the optical capabilities of the PACE, but also on traditional semiconductors and the software solutions that connect the two. The company therefore describes itself as a supplier of hardware and software; the company has also designed algorithms specifically designed to solve some of today's most fundamental computing problems in a photonic environment. PACE isn't very versatile when it comes to what exact workloads it can perform. In this perspective, we can classify it as a kind of ASIC (Application-Specific Integrated Circuit): it does very few things (or a single thing) very, very well.
Photo Credit: Lightenlligence However, PACE achieves the coveted specialization through a further field of information technology, which not only makes the system faster, but also makes it incredibly more efficient. While traditional semiconductor systems have the problem of excess heat caused by the passage of current through nanometer-level elements at very high frequencies, the photonic system processes its workloads with zero ohmic heating: there is no heat produced by the current resistance. . Instead, it's all about the light.
Lightelligence is based on its CEO's PhD. This is because when "Deep Learning with Coherent Nanophotonic Circuits" was published in Nature in 2017, Lightelligence CEO and founder Yichen Chen had already envisioned a path for optical circuits to be at the forefront of computer-based Machine Learning efforts. By 2020, the company had already received $ 100 million in funding and employed approximately 150 employees. A year later, Lightspeed made a product that it says is "hundreds of times faster than a typical processing unit, such as the NVIDIA RTX 3080". 350 times faster, to be clear.
PACE's debut aims to attract enough capital to comfortably achieve its goal of bringing an AI accelerator pilot product to market in 2022. This is still just an ambitious goal in the company's vision, however, its goal is to develop and deploy a photonics-based mass market hardware solution as early as 2023, targeting the Cloud AI, Finance and Retail markets.
