Kria ODM ecosystem accelerates time to market for adaptive computing

Building on the successful reception of the Kria range of adaptive system on modules (SoMs), AMD has announced an ecosystem program designed to deliver production-grade, fully functional Kria SOM-based products that help customers get to market faster without the need for dedicated chip-design resources.

Developers can enjoy a turnkey experience with a full software stack and application support backed by established partners, said AMD.

Currently available ODM partner products developed via the Kria ODM Ecosystem, and powered by the K26 SoM, include, the Ectron IIoT edge gateway, the Optomotive industrial smart camera and the VVDN edge AI appliance.

The Ectron IIoT edge gateway allows users to deploy edge analytics and advanced machine control to optimise big data from smart factories.

The Optomotive industrial smart camera is a programmable, high-speed, industrial camera which can be added to system and used for laser triangulation, motion capture, industrial process automation and industrial quality control.

The VVDN edge AI appliance can be used to deploy an eight-channel AI edge appliance (the Kria K26 SoM and Kinara Ara-1 processors) with support for high-performance vision AI applications at a fraction of the cost of competitive GPU options, claimed AMD. It has application potential in smart city and retail applications.

Initially targeting smart cities, security, retail, industrial IoT and machine vision applications, these and other Kria SoM-based solutions developed by AMD’s industry partners will allow users to take advantage of the processing power and adaptability of Kria SoMs in pre-built, deployment-ready platforms that can quickly be customized, branded and taken to market. 

Featuring adaptive computing with software and hardware re-programmability, users can quickly add features over the air and optimise system performance according to each task’s needs. AMD continued that this provides unparalleled flexibility. 

https://www.xilinx.com/

> Read More

Stroll on – an AI pram rocks baby to sleep

The sensation of CES 2023, the Ella smart stroller, presented by Canadian start-up Glüxkind, is powered by Nvidia Jetson Edge AI.

There is a rock-a-baby feature and a push and brake assist allowing a parent or guardian to hold the occupant while the pram navigates itself.

Just as modern cars are equipped with systems that ensure they do no roll backward when they stop on a hill, Glüxkind founder and CEO, Kevin Huang envisioned a pram (or stroller) that would do the same.

Glüxkind is a member of Nvidia Inception, a free program designed to help start ups evolve faster by providing access to technology and Nvidia experts, as well as opportunities to connect with venture capitalists and co-marketing support.

The Ella pram is powered by Nvidia’s Jetson edge AI platform to power the AI stack.

Jetson enables the pram / stroller to use computer vision to map its surroundings, using Jetson’s GPU and CPU to process and execute pathfinding.

This means that when the child is not in the stroller, parents can activate Ella’s intelligent hands-free strolling mode. It is able to follow someone holding the child, so parents do not have to negotiate their precious bundle while steering the stroller. It also ensure that it does not roll away when the carer’s releases their grip. At the same time the intelligent pram is not designed to operate independently. It has adaptive push and brake assistance for effortless walking no matter the terrain. It can navigate uphill, downhill or even when fully loaded with shopping or toys and baby’s changing equipment.

A Rock-My-Baby mode helps send infants to sleep and there is also built-in white noise playback.

“We’re trying to make it so the technology we’re building is augmentative to the parents’ experience to make parenting easier and safer,” Huang said.

Glüxkind was named a CES 2023 Innovation Awards Honoree for the AI-powered Ella smart stroller. 

http://www.nvidia.com

> Read More

Virtual I/O expansion module helps deploy AI, says Innodisk

The burgeoning AI and AIoT markets could result in 29 billion IoT edge devices by 2030, reports Innodisk. Such a large volume of edge devices means it is necessary to consider not only remote management, but also how to deploy devices more effectively in order to efficiently implement applications. In response, the company has introduced the InnoEx Virtual I/O expansion module, which helps global industries efficiently deploy various AI smart applications through the integration of software and hardware using virtual I/O expansion technology. 

The integration helps enterprises effectively control computer equipment budgets and solve the issue of complicated circuits that cannot be cut and arranged, said Innodisk. It can be extended through virtual I/O technology using only a single computer system and a single network line, and can be connected with the InnoEx modules in series to flexibly expand the number and distance of terminal devices.

Miniaturisation in the computer world means the space for I/O expansion slots has been greatly reduced. If integrators want to increase the number of I/O expansion devices, they often have to redesign the motherboard, which can be time-consuming and not always economical. Embedded expansion cards by Innodisk can solve these issues but not all motherboards can use expansion cards. Innodisk has therefore launched InnoEx virtual I/O expansion. It allows for a single route to connect a single computer system and functions like a multi-function transfer port, effectively expanding the number and distance of external devices through virtual signals. 

Benefits include high system compatibility and it does not occupy the I/O slots on the original computer system. Another advantage is that it does not require changes to the mechanism design. It can also be integrated into the existing network architecture to achieve borderless I/O expansion through networking.

The InnoEx virtual I/O expansion module supports interfaces such as HDMI and USB.  It is suitable for applications such as smart retail, smart image recognition, digital display, and AGV smart handling. It can also realise various industrial applications and advanced AI intelligent applications.

Innodisk has also announced that it will launch Serial and CANbus versions in the second quarter of 2023.

https://www.innodisk.com

> Read More

Intel optimises fourth general Xeon Scalable processors for data centre power

The fourth generation Intel Xeon Scalable processors are Intel’s most sustainable data centre processors, said the company. They have a range of features for optimising power and performance, to help achieve customers’ sustainability goals, said Intel.

Code-named Sapphire Rapids, the Xeon processors are claimed to have the most built-in accelerators of any CPU available today, making them suitable for AI, analytics, networking, security, storage and high performance computing (HPC) applications.

Compared to earlier generations, they increase performance per Watt efficiency by a factor of 2.9, using built-in accelerators, to save up to 70W power per CPU in optimised power mode with minimal performance loss, said Intel. This equates to between 52 and 66 per cent low total cost of ownership (TCO).

The number of built-in accelerators means that the fourth generation Xeon processor needs less additional discrete acceleration.

The new optimised power mode can deliver up to 20 per cent socket power savings with a less than five per cent performance impact for selected workloads. Innovations in air and liquid cooling reduce total data centre energy consumption, added Intel. 

The manufacturing process of the processors uses 90 per cent or more renewable electricity at Intel sites with water reclamation facilities.

In AI, and compared to the previous generation, the processors increase PyTorch real time inference and training performance with built-in Intel Advanced Matrix Extension (Intel AMX) accelerators. 

The low latency network and edge workloads mean that the fourth generation Xeon processors are “a critical part of the foundation driving a more software-defined future for industries ranging from telecommunications and retail to manufacturing and smart cities,” said Intel. For 5G core workloads, the built-in accelerators help increase throughput and decrease latency, while the processors’ power management enhances both the responsiveness and the efficiency of the platform, Intel added. 

Compared to earlier generations, they deliver up to twice the virtualised radio access network (vRAN) capacity without increasing power consumption. This enables communications service providers to double the performance-per-Watt to meet critical performance, scaling and energy efficiency needs.

The latest Xeon processor combines up to four Intel 7-built tiles on a single package, connected using Intel EMIB (embedded multi-die interconnect bridge) packaging technology. It delivers increased memory bandwidth with DDR5, increased I/O bandwidth with PCIe5.0 and Compute Express Link (CXL) 1.1 interconnect.

http://www.intel.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