IMG CXT is “most advanced ray tracing GPU” says Imagination

With the capability to bring desktop-quality visuals to mobile applications, Imagination Technologies announces the IMG CXT GPU IP which uses Imagination’s PowerVR Photon ray tracing architecture. It has up to 50 per cent more compute and geometry performance per core compared to BXT.

By adding Photon hardware ray tracing, the IMG CXT advances GPU IP, says Imagination, delivering performance for games and other graphical use cases. The Photon ray tracing architecture brings desktop-quality visuals to mobile and embedded applications, says the company. It imitates how light behaves in the real world, making it possible to create 3D scenes that are “near indistinguishable from real life”. Ray tracing is relatively recent to desktop PCs and gaming consoles, but IMG CXT’s Photon architecture allows mobile developers and gamers to take advantage of its rendering technology with full hardware acceleration. Photon brings ray tracing to mobile, gaming, AR, desktop, data centre, cloud and automotive designs.

Imagination has identified six levels of ray tracing (from Level 0 to Level 5). The Photon architecture inside IMG CXT is Level 4, enhancing ray tracing performance and efficiency to deliver a desktop-quality experience for mobile gamers and developers.

The IMG CXT-48-1536 RT3 core features three instances of the ray acceleration cluster (RAC), offering up to a total of 1.3Gray per second. This delivers photorealistic ray traced shadows, reflections, global illumination, and ambient occlusion, with high frame rates, in a mobile power budget, says Imagination.

For rasterised graphics performance, IMG CXT has 50 per cent more compute, texturing and geometry performance than Imagination’s previous GPU IP. Its low power superscalar architecture delivers high performance at low clock frequencies for exceptional FPS/W efficiency, while Imagination Image Compression (IMGIC) reduces bandwidth requirements.

The Photon architecture can be scaled to cloud, data centre and PC markets using Imagination’s multi-core technology. This can generate up to 9TFLOPS of FP32 rasterised performance and over 7.8Gray per second of ray tracing performance, offering up to 2.5 times greater power efficiency than Level 2 or 3 RTLS ray tracing.

IMG CXT can also be used for automotive human machine interface (HMI) platforms. It enables the photo-realistic representation of vehicles for surround view and dense urban areas for sat-navs, as well as mixed-reality heads up displays (HUDs), digital speedometers and other information displays. It can also deliver premium in-car gaming and entertainment experiences while occupants wait for the vehicle to charge.

The PowerVR Photon architecture features the RAC, a low power, dedicated hardware GPU block which accelerates and offloads more of the ray tracing computations from the shader cores compared to less-efficient Level 2 RTLS architectures.

The RAC consists of the ray store, ray task scheduler and coherency gatherer and is closely coupled to two 128-wide unified shading clusters (USCs) featuring high-speed dedicated data paths for the most efficient and lowest power ray traced deployment. The ray store keeps ray data structures on-chip during processing, providing high-bandwidth read and write access to all units in the RAC, avoiding slowdowns or power increases from storing or reading ray data to DRAM. The ray task scheduler offloads the shader clusters, deploying and tracking ray workloads with dedicated hardware, keeping ray throughput high and power consumption low. The coherency gatherer unit analyses all rays in flight and bundles rays from across the scene into coherent groups enabling them to be processed with much greater efficiency.

These features combine to deliver greater and more consistent ray tracing performance for developers than current solutions on the market, says Imagination, for advanced effects when rendering complex surfaces such as cars, characters, and terrain, to create more detailed environments than previously possible.

http://www.imaginationtech.com

> Read More

Tool accelerates edge-native AI development

To simplify and accelerate the design, development, testing and deployment of smart systems, MicroAI has introduced Launchpad, a quick start development and deployment tool. It is designed to run embedded MicroAI software on microcontrollers  and microprocessors in edge and endpoint devices.

The tool handles customers with SIMs around the world and provides a flexible way to manage and reconfigure device profiles. Launchpad allows engineers to customise dashboards, including account creation, authentication, mobile SIM or LoRaWAN connectivity activation, credit card billing for global SIM connectivity with MicroAI’s embedded software libraries.

“MicroAI’s goal is to democratise the development of smart machines for all organisations across any industry,” said MicroAI CEO Yasser Khan. “Regardless of industry or product, building . . .  smart device includes creating an edge AI model, but also integrating connectivity and cloud resources, as well as device activation and management.”

MicroAI’s embedded software, AtomML, enables OEMs to deploy personalised, edge-native AI models, without needing to develop static edge-AI models first in a cloud or laptop and then port them to the embedded device. Instead, MicroAI AtomML moves the training and inferencing directly to the embedded device. Launchpad then simplifies and reduces the time and cost to integrate the microcontrollers and microprocessors into an edge device, which can be tested and scaled to proof of concept for mass deployment.

MicroAI Launchpad can be white labelled for use by semiconductor companies, OEMs  and service providers. Semiconductor companies that offer stock keeping units (SKUs) with MicroAI embedded AI software can leverage its end-to-end device management and deliver it to their customers to help expedite design, development, testing and deployment. OEMs directly engaged with MicroAI benefit from Launchpad’s flexibility to evaluate various hardware, software, and cloud solutions before finalising a deployment model. For IoT service providers, Launchpad will provide device certification, connectivity, and deployment for a deeper insight view of connected devices on networks.

MicroAI is based in Dallas, Texas, USA and specialises in edge-native artificial intelligence (AI) and machine learning (ML) products. The company is personalising AI for connected machines, edge devices, and critical assets by embedding its proprietary edge-native AI technology directly onto microcontrollers and microprocessors within edge endpoints. This enables device-specific and more accurate AI modelling for edge and endpoint cyber security, advanced predictive maintenance, IoT performance optimisation and significant improvements in overall equipment effectiveness (OEE). The company’s mission is to democratise edge-native AI for all connected, smart devices by reducing the complexity, time, and cost to design, develop, and deploy embedded, edge-native AI.

http://www.micro.ai

> Read More

IAR Embedded Workbench supports Arm Cortex-55 core for AI

Support for the latest Arm Cortex-M55 processor has been announced by IAR Systems. The latest version of IAR Embedded Workbench for Arm adds support for the Arm Cortex-M55 processor. In addition, version 9.20 of the toolchain includes support for latest microcontroller devices from several semiconductor vendors.

The Arm Cortex-M55 processor is Arm’s AI-capable Cortex-M processor and the first to feature Arm Helium technology, M-Profile Vector Extension (MVE). It is characterised by energy-efficient digital signal processing (DSP) and machine learning (ML) capabilities. The IAR Embedded Workbench for Arm toolchain delivers powerful optimisation capabilities to assist developers in getting the most out of the performance of the microcontroller while maintaining energy efficiency. To ensure code quality, code analysis tools are completely integrated with IAR Embedded Workbench.

The support for Cortex-M55 is intended to help early development based on the core in the ecosystem. 

In a simultaneous announcement, the company announced that the latest IAR Build Tools for Arm provides support for Linux and Windows installations, which enables implementation in cross platform-based frameworks and large-scale deployments of critical software building and testing. 

IAR Systems says it supplies future-proof software tools and services for embedded development, enabling companies worldwide to create the products of today and the innovations of tomorrow. Since 1983, IAR Systems’ solutions have been used in the development of over one million embedded applications. 

The company is headquartered in Uppsala, Sweden and has sales and support offices all over the world. In 2018, Secure Thingz, the global domain expert in device security, embedded systems, and lifecycle management, became part of IAR Systems Group AB. 

http://www.iar.com

> Read More

Face recognition uses 3D SLM camera and NXP’s i.MX RT117F crossover MCU 

A 3D face recognition access control systems uses  NXP Semiconductor’s crossover MCU with a 3D camera to enable faster secure face recognition under challenging lighting conditions.

The 3D structured light module (SLM) camera is combined with the i.MX RT117F crossover MCU and is believed to be the first time a 3D SLM camera has been combined with an MCU to deliver the performance and security of 3D face recognition at the edge. It therefore removes the need to use an expensive and power-hungry Linux implementation on a microprocessor which is conventionally the case with high-performance 3D cameras, reports NXP.

The i.MX RT117F MCU is part of the i.MX RT1170 family of crossover MCUs. It is based on an Arm Cortex-M7 CPU with 2Mbyte of on-chip SRAM, running at up to 1GHz. 

The turnkey system is the latest EdgeReady solution from NXP. It enables developers of smart locks and other access control systems to add machine learning-based secure face recognition quickly and easily to smart home and smart building products. Reliable 3D face recognition can be achieved in indoor and outdoor applications, across varied lighting conditions, including bright sunlight, dim night light, or other difficult lighting conditions that are challenging for traditional face recognition systems.

The use of a 3D SLM camera enables advanced liveness detection, helping distinguish a real person from spoofing techniques, such as a photograph, imitator mask or a 3D model, to prevent unauthorised access.

The i.MX RT117F uses an advanced machine learning model as part of NXP’s eIQ machine learning software running on its CPU core.

 Advanced liveness detection and face recognition are performed locally at the edge, making it possible for personal biometric data to remain on the device. This helps address consumer privacy concerns, while also eliminating the latency associated with cloud-based solutions.

The development kit for this 3D face recognition system, the SLN-VIZN3D-IOT, will be available later in November from NXP and authorised distributors.

The i.MX RT117F includes a license to use the NXP 3D face recognition software development kit (SDK), and is available in consumer, industrial and automotive temperature grades.

http://www.nxp.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