Neurocomputing Lab Creates First Russian Neural Network Processor
February 7, 2019 | MIPTEstimated reading time: 2 minutes
Engineers and researchers working under a project of the Russian Foundation for Advanced Research (FAR) — a government agency for defense research — has demonstrated the first Russian-made specialized neural network processor for energy efficient execution of machine learning algorithms based on deep neural networks. These algorithms are widely used to automatically solve tasks in image and speech recognition, self-driving car control, and elsewhere. The prime contractor of the project is the Neurocomputing Systems Lab of the Moscow Institute of Physics and Technology.
Machine learning algorithms that rely on deep neural networks require considerable computing power, and they are typically run using graphics accelerators. The latter consume way too much power to be of use on a mobile platform, such as small unmanned aerial or ground vehicles or smart cameras. This is why researchers worldwide are pursuing the notion of specialized neural network tensor processors. By leveraging certain computing capabilities of neural networks, tensor processing units can achieve a much higher energy efficiency, compared with graphics accelerators.
MIPT has demonstrated NCore, a neural network processor created under the 65-nanometer design standard. Although the processor was manufactured outside Russia, the required technology is currently being introduced at a Mikron facility in Zelenograd, a suburb of Moscow. NCore achieves an energy efficiency comparable to that of graphics accelerators complying with the more advanced 28-nanometer design standard or even stricter guidelines.
According to Alexey Zablotskiy, who heads the FAR project group, the architecture can be scaled down to create Russian neural network processors under “thinner” design standards. This would ensure scientific and technological parity in specialized tensor processing unit design for machine learning algorithms, Zablotskiy added.
“By now we have developed the basic elements: processor cores, interfaces, and the essential software,” said Dmitriy Negrov from the MIPT Neurocomputing Systems Lab. “Our software for algorithm translation enables us to convert a neural network to launch it with our system using popular machine learning frameworks.”
“We already have a system prototype in silicon, manufactured using the 65-nanometer technology, which is fairly coarse by modern standards. Eventually we intend to use our cores as components for crystal-based systems,” the researcher added.
An important part of the project is developing an automated translator for machine learning algorithms developed using the popular frameworks TensorFlow, Caffe, and Keras. This unification will enable users to work in familiar environments.
NCore testing with prospective consumer participation is scheduled for spring 2019. Following this, an improved version of the processor will be created.
Project co-contractors alongside the MIPT lab include the State Research Institute for Aviation Systems and the Bruk Institute of Electronic Control Machines. The project is supported by the Russian Ministry of Education and Science
Suggested Items
Elementary, Mr. Watson: Ensuring Design Integrity
03/28/2024 | John Watson -- Column: Elementary, Mr. WatsonBack in February, many of us watched the "Big Game." It reminded me of the saying, “It's not how you start that is important, but rather how you finish." It is perfectly okay when you are talking about sports, you get off to a bad first half and need to recover in the second half. However, when it comes to PCB design, this is not a good practice. If things start badly, they usually don't recover. They continue down that same path, costing more money and losing design time.
Arrow Electronics Launches Intelligent Vision Ecosystem
03/27/2024 | BUSINESS WIREArrow Electronics, Inc. is utilizing the onsemi Imager Access System (IAS) module standard for developing intelligent vision solutions for use in robotics, machine vision, commercial cameras and other uses.
Dymax Will Exhibit Light-Cure Solutions for Today’s Electronics at IPC APEX 2024
03/26/2024 | DymaxDymax, a leading manufacturer of rapid and light-curing materials and equipment, will exhibit at the IPC APEX EXPO 2024 in Anaheim, CA, April 9-11.
Synopsys Announces New AI-Driven EDA, IP and Systems Design Solutions At SNUG Silicon Valley
03/25/2024 | PRNewswireSynopsys, Inc. kicked off its annual flagship Synopsys User Group (SNUG) conference in Silicon Valley at the Santa Clara Convention Center with a keynote presentation by Synopsys president and CEO Sassine Ghazi. Ghazi discussed the unprecedented innovation opportunities and challenges that technology R&D teams face in this era of pervasive intelligence.
New North American Product Family Release from Smart Factory Design
03/22/2024 | Smart Factory DesignSmart Factory Design is pleased to announce a new system technology and supply chain model for providing Zero Liquid Discharge (ZLD) Wastewater Recycling Systems to the North American market. Smart Factory Design is an international public benefit company focused on environmental sustainability and autonomous processes for the Printed Circuit Board, IC Substrate, and General Metal Finishing industries.