Software accelerates artificial intelligence workloads

Artificial intelligence (AI) –specific tiles make up a library used by eSilicon in its 7nm FinFET newASIC AI IP platform. The AI Accelerator tiles can be configured to support an AI algorithm and map high-level AI workloads to the neuASIC platform and estimate power, performance, area (PPA) for the algorithm in the silicon implementation.

The combination of neuASIC IP and the AI Accelerator software allows designers to explore architectures to ensure the design will be within the target specifications. eSilicon explains that this approach supports changes to the algorithm or the package.

AI Accelerator is available in IP Navigator, eSilicon’s IP exploration and evaluation tool, at no charge. 

eSilicon provides complex FinFET ASICs, market-specific IP platforms and advanced 2.5D packaging. The ASIC-proven, differentiating IP includes highly configurable 7nm 56G/112G SerDes plus networking-optimised 16/14/7nm FinFET IP platforms featuring HBM2 (high bandwidth memory) PHY (physical layer), ternary content-addressable memory (TCAM), specialised memory compilers and I/O libraries. 

The neuASIC platform provides AI-specific IP and a modular design methodology to create adaptable, efficient AI ASICs. eSilicon serves the high-bandwidth networking, high-performance computing, AI and 5G infrastructure markets. 

http://www.esilicon.com
> Read More

Speech inference is optimised for Intel FPGA PAC to save power demands

To save electricity consumption and to reduce data centre infrastructure, Myrtle announces that its artificial intelligence (AI) can run on the new, high-performance Intel FPGA Programmable Acceleration Card (Intel FPGA PAC) D5005 accelerator. The result is to reduce costs and remove growth constraints for businesses offering speech services such as transcription, translation, synthesis or voice assistance in on-premise or cloud-based data centres, says the AI specialist.

Intel and Myrtle have worked together to optimise a recurrent neural network (RNN) for speech inference on the Intel FPGA PAC D5005. The collaboration can run more than 4,000 voice channels concurrently on one FPGA, leading to a six-fold improvement in performance per watt compared with general purpose GPUs with a latency of one 30th that of a GPU, reports Myrtle.

“The industry has to take new approaches to produce machine learning solutions that meet customers’ stringent latency, power and cost constraints”, said Peter Baldwin, CEO, Myrtle. He added that these performance metrics on Intel’s latest PCA will allow customers preserve their investment in hardware as machine learning models evolve.

Myrtle specialises in hardware-software codesign. The quantisation, sparsity and compression of machine learning models has been recognised by the MLPerf consortium. Myrtle dominates MLPerf speech transcription and has open sourced its code to help the industry benchmark new edge and data centre hardware more consistently, says the company.

Myrtle creates high-performance, energy-efficient computing solutions for deep learning inferencing on next-generation data centre hardware. Myrtle’s RNN technology enables companies to cost-efficiently implement and scale speech applications on cloud or on-premise infrastructure.

Myrtle is a partner in Intel’s design solutions network (DSN).

http://www.myrtle.ai

> Read More

Secure flash memory enhances secure data storage in self-driving cars

Macronix’s secure flash memory has been integrated in Nvidia’s next-generation autonomous driving platforms.

The automotive-grade ArmorFlash memory is being used on the Nvidia Drive AGX Xavier and Drive AGX Pegasus autonomous vehicle computing platforms.

The ArmorFlash memory is secure for data storage in the artificial intelligence (AI)-based Level 2+ advanced driver assistance systems (ADAS) through to Level 5 autonomous driving.   

“Our efforts in conjunction with NVIDIA are singularly focused on elevating the security of data in AI-based autonomous driving applications and ultimately, to enhance the safety of drivers,” said Anthony Le, vice president of marketing, Macronix America.

The ArmorFlash memory on the Drive AGX Xavier and Pegasus platforms can provide trusted identification, authentication and encryption features for autonomous driving security requirements.

ArmorFlash offers a combination of mature security technologies, including unique ID, authentication and encryption features. This blend of features enables superior levels of security in a high-density memory device to prevent data from being compromised, claims Macronix.

The ArmorFlash device provides trusted NVM storage of encrypted and integrity-protected assets. The ArmorFlash supports secure communication channel and protocol with the Nvidia Xavier system on a chip (SoC) via cryptographic operations, integrity checks and additional measures against certain security protocol attacks.

The global ADAS market is expected to exceed $67 billion by 2025, fuelled by a compounded annual growth rate of 19 per cent, according to Grand View Research. The research company attributes the growth to increasing government initiatives mandating driver assistance system to lower road accidents and cites expanding adoption of ADAS in small cars as a factor boosting market demand.

http://www.macronix.com

> Read More

Skyline RFID inlays and tags can be used on metal

Rain radio frequency identification (RFID) inlays can be used on metal surfaces for and have a reading range of 6m.

The Skyline Rain RFID inlays and tags have an optimised antenna and spacer-based design. A customised ultra high frequency (UHF) inlay is based on NXP’s UCODE 7xm IC with 448-bit EPC memory and extended user memory of 2kbits. The resulting transponder is then folded and applied around a synthetic spacer developed and provided by identytag of Bad Berleburg, Germany.

The advanced antenna design, the IC’s long read range and reliable operation in dense reader and noisy environments through high interference rejection, as well as  optimised spacer material result in an on-metal read range in a compact tag, with a die-cut size of 54 x 25 x 1.8mm.

The inlay is permanently attached to the spacer and a layer of strong and resilient RA-33 adhesive is applied. According to Smartrac, this provides “excellent adhesion” to a range of surfaces. As a finished tag, Skyline’s surface is printable with thermal transfer printers.

The RFID tag can be used for tracking metallic assets, items and components in industrial environments such as automotive, mechanical engineering and aviation. Smartrac’s Skyline inlays and tags comply with VDA recommendations for the automotive industry and supported by the leading automation companies globally.

Smartrac and identytag completed initial product volumes and will be ramping up production in the second half of the year 2019.

Smartrac provides both ready-made and customised products. It makes products smart and enables businesses to digitise, identify, authenticate, track and complement products. Products are used in a wide array of applications such as animal identification, automation, automotive, brand protection, customer experience, industry, library and media management, logistics, retail and supply chain management.

Based in Amsterdam, the Netherlands, Smartac has research and development centres, production and a sales network, complemented by the IoT platform, Smart Cosmos. Smartrac embeds intelligence into physical products for an ecosystem of connected things. The company has also received ARC Quality Certification from Auburn University’s RFID Lab for the design and manufacturing of its RFID inlays.

http://www.smartrac-group.com

> Read More

About Smart Cities

This news story is brought to you by smartcitieselectronics.com, the specialist site dedicated to delivering information about what’s new in the Smart City Electronics industry, with daily news updates, new products and industry news. To stay up-to-date, register to receive our weekly newsletters and keep yourself informed on the latest technology news and new products from around the globe. Simply click this link to register here: Smart Cities Registration