Story image to build supercomputers for edge devices

20 May 2019
Facebook is making a bold entrance into the computing space announcing it is building the world's first supercomputer specifically designed for the real-time needs of autonomous systems. Built to increase computation speed while reducing power consumption, Quadric's supercomputer can lead to increased safety and performance in better-designed machines. 

While innovative developers are reinventing existing machines to be more efficient and responsive across multiple trillion-dollar industries like construction, transportation, agriculture, and warehousing, the underlying technology supporting these machines either lacks the performance or solves only a small part of the full application.

These new machines incorporate autonomous functions that require near instantaneous processing speed and responsiveness "on the edge" (at the machine level, not in the cloud).

Quadric's technology supposedly enables both high-performance computing and artificial intelligence in a single unified architecture. Built around the Quadric Processor, Quadric's supercomputer is plug-and-play, enabling innovators to plug in their sensor set and build their entire application to support near instantaneous decision making. Early testing of Quadric's system has shown up to 100x lower latency and a 90% reduction in power consumption.

The funding round is led by strategic investor DENSO Group Company, NSITEXE, Inc. DENSO CORPORATION is a tier 1 automotive company, who will also be one of Quadric's customers for future electronic systems in all levels of autonomous driving solutions.

"Instead of building a heterogeneous system with CPUs/GPUs/FPGAs & AI chips and dealing with that complexity of software and hardware integration, Quadric has the technology to handle most of the application in a single unified hardware architecture and software model," said Hideki Sugimoto NSITEXE and DENSO CORPORATION. "The integrated approach achieves better latencies that improve the safety of automotive ADAS implementations."

"We're developers ourselves and have seen the shortcomings of existing technologies for real-time edge computing -- we're building the Quadric platform to address this and unlock new capabilities that will accelerate the development of autonomous systems," said CEO Veerbhan Kheterpal.

"Quadric is the next step in the virtuous cycle of better platforms enabling better algorithms, leading to better app engagement that in turn drives better innovation."