Provide the fastest response and the most energy-efficient neural network acceleration function for next-generation automotive application scenarios
London, United Kingdom, November 12, 2020-Imagination Technologies announced the launch of IMG Series4, a new generation of neural network accelerator (NNA) products for advanced driver assistance systems (ADAS) and autonomous driving applications . Series4 can provide a powerful boost to leading auto industry disruptors, Tier 1 suppliers, OEMs and SoC manufacturers.
Series4 has a new multi-core architecture that can provide 600 TOPS (trillion operations per second) or even higher ultra-high performance, and can provide low bandwidth and extremely low latency for large neural network workloads.
The automobile industry is on the cusp of change, and new application scenarios such as autonomous vehicles and autonomous taxis have put forward higher requirements for the performance level of artificial intelligence (AI). To this end, Imagination has collaborated with leading companies and innovators in the automotive industry and other industries that value functional safety.
Series4 has begun to provide authorization and will be fully available on the market in December 2020.
Imagination's low-power NNA architecture is designed to perform complete network inference while meeting functional safety requirements. It can perform multiple operations at once to maximize performance per watt and provide industry-leading energy efficiency.
Series4 has the following characteristics:
Multi-core scalability and flexibility: Multi-core architecture supports flexible distribution and synchronization of workloads among multiple cores. Imagination's software provides fine-grained control capabilities and increases flexibility by batching, splitting and scheduling multiple workloads, and can now be used on any number of cores. Series4 can be configured with 2, 4, 6, or 8 cores for each cluster.
Ultra high performance: Each single core of Series4 can provide 12.5 TOPS performance with less than one watt of power consumption. For example, an 8-core cluster can provide 100 TOPS of computing power, so a solution with 6 8-core clusters can provide 600 TOPS of computing power. In AI reasoning, the performance of Series4 NNA is more than 20 times faster than embedded GPU and 1000 times faster than embedded CPU.
Ultra-low latency: By combining multiple single cores into a 2-core, 4-core, 6-core or 8-core multi-core cluster, all cores can work together to process one task in parallel, reducing processing delay and shortening response time. For example, for an 8-core cluster, ideally the latency will be reduced to 1/8 of that of a single-core independent execution.
Save a lot of bandwidth: Imagination's Tensor Tiling (Imagination's Tensor Tiling, ITT) is a patent-pending technology and a new feature in Series4. It can tiling computing tasks and make full use of on-chip storage to improve data processing Efficiency and save bandwidth for accessing external storage. ITT takes advantage of the dependence of local data to store intermediate data in on-chip memory, which can minimize the transfer of data to external memory, thereby reducing bandwidth by up to 90%. ITT is a scalable algorithm that has significant advantages on networks with large amounts of input data.
Vehicle-level safety: Series4 includes IP-level safety functions and the design process complies with ISO 26262 standards, which can help customers obtain ISO 26262 certification. ISO 26262 is an industry safety standard designed to address the risks of automotive electronic products. Series4 can safely perform neural network inferences without affecting performance. The hardware security mechanism can protect the compiled network, network execution and data processing pipeline.
James Hodgson, Principal Analyst of Smart Mobility and Automotive at ABI Research, said: "Although we are still predicting that the demand for ADAS will triple by 2027, the automotive industry has already set its sights on fully autonomous vehicles and autonomous vehicles. Driving a taxi. In the process of evolving from L2 and L3 ADAS to L4 and L5 fully automated driving, the wide application of neural networks will be a crucial factor. These systems will have to deal with hundreds of complex scenarios, from Data is extracted from a large number of sensors such as multiple cameras and lidar to realize solutions such as automatic valet parking, intersection management and safe navigation in complex urban environments. The combination of high performance, low latency and high energy efficiency will achieve a high degree of autonomous driving The key to it."
Andrew Grant, senior director of artificial intelligence business at Imagination Technologies, said: "We believe Series4 NNA will become the industry standard platform for the development of advanced driver assistance systems and autonomous vehicles. Some innovators are already building chips that support next-generation ADAS functions and autonomous vehicles. It’s time for any company or R&D team that wants to play an important role in the automotive industry to integrate this technology into their platform."
For more information about Series4 NNA, please recognize the QR code below to watch our themed conference.
END
Recommended reading
Imagination Technologies is a UK-based company dedicated to the development of chips and software intellectual property (IP). Products based on Imagination IP have been used in the phones, cars, homes and workplaces of billions of people around the world. For more cutting-edge technology information such as the Internet of Things, smart wear, communications, automotive electronics, graphics and image development, please pay attention to Imagination Tech!
Long press to identify QR code
Follow us