ICs & Semiconductors Archives - Page 68 of 235

Imagination says IMG CXM is the smallest GPU to support HDR user interfaces

Wed, May 24, 2023

Seamless visual experiences for cost-sensitive consumer devices are assured with the IMG CXM GPU of RISC-V compatible cores, said Imagination Technologies. The range includes what is claimed to be the smallest GPU to support HDR user interfaces natively.

The IMG CXM cores can lower the cost of DTV and other consumer devices in smart homes, for example, said Imagination. Another driver is content providers seeking to integrate 4K and HDR features to enrich content, continued the company.

There are three new configurations which extend the range of performance options already available in Imagination’s GPU family of products for the consumer space. Imagination has released the CXM-2-64, claimed to be the smallest GPU to support native HDR applications. It is suitable for wearable devices, SmartHubs or mainstream set-top boxes.

The CXM-4-64 is suitable for integrating into SmartHubs, set-top boxes or mainstream DTVs and the CXM-4-128 is a performance dense option for premium DTVs, advised Imagination.

The company has boosted the performance density of the IMG BXE and BXM range of GPUs for the CXM GPUs and added native support for HDR.

The CXM GPUs support 10bits RGBA / YUV to deliver a HDR graphical user interface with images with less visible banding. To smooth the outline of texts and images, they employ 4xMSAA (Microsoft Active Accessibility, an application programming interface (API) for user interface accessibility).

TFBCv2, the new generation of Imagination’s Tiny Frame Buffer Compression, delivers higher quality lossy / lossless compression and an additional compression level (37.5%) for improved design flexibility.

The IMG CXM GPU range boasts nearly a 50 per cent uplift in performance density compared to the IMG BXM range.

The IMG CXM is supported by software that supports APIs, including Vulkan 1.3, and has been optimised for leading CPU architectures including Arm and RISC-V application processors.

Dr. Charlie Su, CTO and president of Andes Technology said: “The RISC-V ecosystem is growing rapidly. To continue its growth and showcase the many possible ways it can be deployed, we partnered with Imagination to provide a quick and easy path to validated GPU and CPU IP blocks that can reduce SoC design time, risk, and cost for our customers. With Imagination’s flexibly designed GPU, and our AndesCore high performance, low power RISC-V CPU, we are able to satisfy the requirements for display-oriented SoC in a short time and generate the optimum configuration.”

James Chapman, chief product officer, Imagination Technologies commented the CXM GPUs will “transform user experiences” and he expects to see the GPUs deployed in a diverse range of applications from wearables to premium 8K DTVs.

Imagination will be demonstrating TFBCv2 at the 2023 RISC-V Con Shanghai and 2023 RISC-V Con Beijing events, hosted by Andes Technology.

https://www.imaginationtech.com/

> Read More

Intel’s Agilex 7 FPGAs have PCIe 5.0 and CXL capabilities

Tue, May 23, 2023

Believed to be the first FPGA with PCIe 5.0 and CXL (Compute Express Link) inteconnect technology capabilities, the Agilex 7 FPGAs is shipping now with the R-Tile chiplet. Hardware support for the fast communication and interconnect standards will serve the increasing bandwidth requirements of cloud processing and is also claimed to be the only FPGA with hard IP supporting these interfaces.

The Agilex 7 FPGA with R-Tile addresses the need for fast, flexible devices to accommodate new connectivity and processing models from the edge to the cloud and increasing bandwidth requirements across markets. The FPGA has the high bandwidth interfaces and flexible programmable logic needed to address these requirements.

Market adoption of FPGA accelerators has steadily increased in recent years and with the rollout of R-Tile-equipped FPGAs, reported Intel. FPGA accelerators can offload tasks from the host CPU, freeing up CPU cores and reducing the total power consumed, enabling total cost of ownership (TCO) savings. Cloud providers using FPGA accelerators can support more users and generate more licensing dollars from the newly available CPU cores, advised Intel.

Intel Agilex 7 FPGAs are built with a heterogeneous multi-die architecture, with an FPGA fabric chiplet in the centre connected to transceiver chiplets via Intel’s embedded multi-die interconnect bridge (EMIB) technology. Each chiplet, or tile, is a small IC die containing a well-defined subset of hardened functionality. These chiplets enable a cost-effective approach to the in-package, high-density interconnect of heterogeneous chips addressing a broad array of applications with tailored, flexible devices and realise connectivity topologies within a single device that previously would have required multiple devices.

Many of the Intel Agilex 7 FPGA package combinations include the R-Tile chiplet, designed to support industry-leading bandwidth when connecting to high performance CPUs. The R-Tile chiplet combines hard IP blocks and soft IP code for PCIe 5.0 x16 and CXL 1.1 / 2.0, providing a high degree of flexibility across networking, cloud, data centre, high performance computing, for example.

R-Tile-based Intel Agilex 7 FPGAs are now in production.

The production qualification of R-Tile triggers the production release sequence for seven device densities across four different packages within the Intel Agilex 7 FPGA I-Series devices enabling customers to leverage Intel Agilex 7 FPGA fabric performance / per Watt leadership on new designs.

Built on the Intel 10nm process technology, both the Intel Agilex 7 FPGA programmable logic and R-Tile chiplet leverage Intel’s robust supply chain with advanced manufacturing and test capabilities to deliver production solutions to standard lead times. Additional device density and package options will reach production once the Intel Agilex 7 FPGAs M-Series with R-Tile transition from sampling to production.

http://www.intel.com

> Read More

Intel updates Flex GPU for cloud gaming and AI

Fri, May 19, 2023

Aoption of Intel’s Flex series GPU continues to grow, reported the company, leading intended for flexible, general-purpose graphics processing for the data centre and the intelligent visual cloud, Intel has expanded the Flex GPU’s production-level software capabilities to include new support for Windows cloud gaming, AI inference and digital content creation.

The Flex series GPUs can be used by cloud service providers deploying Windows cloud gaming alongside streaming and media acceleration. They are also used by media studios for high-density streaming and transcoding. Digital content creators are using the platform for fast, real-time rendering using ray tracing hardware acceleration.

The Flex series GPUs address the gaming market which is projected to grow at a CAGR of 42.5 per cent from 2022 to 2028 with the global cloud gaming market estimated to reach about $13.3 billion by 2028. Game service providers must continually innovate to deliver playing experiences to subscribers while operating the most efficient infrastructure possible, advised Intel.

Intel’s new software capabilities for the Intel Flex series GPU enable customers to realise new capabilities and gains across real-world workloads coupled with third Gen and fourth Gen Intel Xeon Scalable processors.

Intel provides a cloud gaming reference stack for Windows to show how to unleash Intel Flex Series for remote gaming, including how to enumerate and retarget multiple game titles to run concurrently on each of the multiple Flex Series adapters and two types of virtual displays. Intel Flex series is optimised for DirectX 9, 11 and 12 render target capture and desktop capture.

Gamestream allows users to stream and play games on devices with high-end graphics across the entertainment, hospitality, media and telecomms markets.

Olivier Lebigot, Gamestream’s CTO, said, “One of our key challenges in cloud gaming is to find the right GPU to increase the overall concurrent users (CCU) number per server, while delivering the best end-user experience. We have been pleased to find that Intel Data Center GPU Flex Series 170 offers a high density, low power solution. During our tests, our reference CCU was improved by nearly 19 per cent compared to our current hardware solution”.

Intel has expanded its Flex series capabilities for AI, including workloads for smart city, library indexing and compliance, AI-guided video enhancement, intelligent traffic management, smart buildings and factories, and retail applications.

Through the Intel AI Analytics toolkit, Intel Flex series GPU supports most common AI frameworks, including TensorFlow and PyTorch. Intel has validated more than 100 AI inference models.

The Intel software stack for visual inference now includes the open source GStreamer multimedia framework, the Intel OpenVINO toolkit inference engine and open source sample implementation of media analytics framework for creating complex visual inference and analytics pipelines.

A prominent open source solution for supporting the advancement, training and validation of autonomous driving systems is CARLA, which utilises the Unreal Engine 4.

In a single 1080p sensor scenario, a single Intel Data Center Flex 170 GPU achieves 56 frames per second (fps), which is 16 per cent faster than the Nvidia A10G (48 fps), 40 per cent faster than the Nvidia GPU Tesla T4 (40 fps) and 60 per cent faster than the AMD Radeon Pro V520 MxGPU (35 fps), reported Intel.

The performance of the Intel Data Center Flex 170 GPU scales accordingly for four 1080p sensors, reaching 23 frames per second, while the Nvidia A10G achieves 19 frames per second.

In digital content creation, real-time rendering is delivered via ray tracing hardware acceleration on the Intel Flex Series through Intel Embree. AI-based denoising can be completed in milliseconds using the latest Intel Open Image Denoise. Further productivity can be gained through the use of a single SYCL codebase for CPU/GPU rendering through oneAPI, said Intel.

Flex Series offers customers a comprehensive graphics solution, an open and full software stack, no licensing fees, and a unified programming model for CPUs and GPUs for performance and productivity via oneAPI.

http://www.Intel.com.

> Read More

Reclocker / redriver devices extend reach with USB 3.2-compatibility

Wed, May 17, 2023

Suitable for automotive and industrial applications, Microchip’s EQCO510 and EQCO5X31 devices offer a two-channel option to engineers wishing to send high-speed data signals up to 15 meters in both directions

While the standard universal serial bus (USB) connection transfers data between two devices, the increase in electronic components in applications across the automotive, industrial and consumer industries has spurred the need for far-reaching USB cabling products. Microchip Technology announces two new reclocker / redriver devices for long range, reliable USB operations in industrial and automotive applications.

The automotive EQCO510 and industrial EQCO5X31 reclocker / redriver devices extend USB coverage up to 15 meters – rather than the standard three meters – for maximum reach and are compatible with the USB 3.2 Generation 1 SuperSpeed protocol.

The USB reclocker / redriver devices can send high-speed data signals at a rate of 5Gbits per second in both directions. The reclocking feature includes a bit-level clock data recovery (CDR) that is used to restore signal timing and prevent jitter accumulation. The redriving feature restores the levels and shape of the signal being driven into the next segment, such as a cable or PCB trace, thus compensating for signal degradation due to cable attenuation, explained Microchip.

“Our customers now have the capability to implement a fast, reliable and long-distance USB connectivity solution of up to 15 meters, extending the three meter standard by five times,” said Matthias Kaestner, corporate vice president of Microchip’s automotive infotainment systems business unit. These devices also reduce board space because of the on-chip clock and small form factor, he added.

The USB devices are equipped with EyeOpen cable compensation at the receiver to automatically adjust for frequency dependent losses in the cable and adjusting the signal strength between 0 and 24dB with 1.0dB steps. The devices also feature MarginLink signal integrity testing, which allows runtime evaluation of the integrity of the whole signal path.

The EQCO510 and EQCO5X31 ICs support shielded twisted pair and coax cables. The devices include an integrated crystal-less CDR, which reduces the need for additional components and overall board space.

Both USB devices are available in a 20-pin, 4.0mm QFN package with wettable flanks. The automotive EQCO510 complies with AEC-Q100 Grade 2 and operates in temperatures ranging from -40 to +105 degrees C.

Target applications are in automotive infotainment systems, data communication modules and real-time video systems. In industrial applications and consumer products they can be used in machine vision, gaming accessories and smart cables.

Microchip offers two evaluation boards for the EQCO5X31, which include the EVB-EQCO5X31 USB Type-C cable extender or the EVB-EQCO5X31 USB Type-C cable repeater to demonstrate use with longer cables.

http://www.microchip.com

> Read More

Imagination says IMG CXM is the smallest GPU to support HDR user interfaces

Intel’s Agilex 7 FPGAs have PCIe 5.0 and CXL capabilities

Intel updates Flex GPU for cloud gaming and AI

Reclocker / redriver devices extend reach with USB 3.2-compatibility

About Smart Cities