The second week of the conference will focus on Xilinx’s diversified product portfolio and terminal market segments, and display the latest products, solutions and industrial collaboration.
On September 8, 2021, Beijing, China (Xilinx adapt) – Xilinx, Inc. (NASDAQ: xlnx), a leading enterprise in adaptive computing, released a series of contents on Xilinx adapt 2021 today, including by alveo ™ New Microsoft database analysis service supported by accelerator card, software development kit (SDK) for real-time video transcoding, KRIA ™ Robot stack, latest Vitis for medical ultrasound ™ Library, and the new zynq ® Rfsoc DFE has been mass produced and supplied to customers all over the world. Xilinx adapt 2021 is a virtual technology conference held by Xilinx, which was held from September 7 to 16, us time. On November 16, 2021, Xilinx adapt China virtual event will be launched soon. As a special content in China, adapt 2021 China will face the Chinese market and bring many wonderful contents by Xilinx local speakers, partners and customers.
Xilinx adapt 2021 includes the company’s top-level keynote speech, as well as the experience of Xilinx partners and customers. In the same period, there will be more than 100 lectures, forums, product training and laboratory projects to help users unlock the value of adaptive computing. The adapt 2021 event will begin this week for six days and is free. Among them, the content from September 7 to 9, us time is for software and hardware developers; From September 14 to 16 U.S. time, the content focused on Xilinx’s extensive terminal market, and the special speech covered aerospace, automobile, data center, industry, medical treatment, wired and wireless services.
Victor Peng, President and CEO of Xilinx, said: “From data center and 5g to robotics, we look forward to Xilinx adapt 2021 to show our new products and tools for diversified end markets to customers and developers. We are happy to share with participants how adaptive computing can help customers achieve great innovations and bring them to market at an unprecedented speed.”
In the second week of Xilinx adapt 2021, the content focuses on the terminal market segment, and the relevant key points are as follows:
Xilinx FPGA and adaptive SOC can bring great value to lidar sensor architecture because they can support advanced signal processing, point cloud preprocessing and point cloud ml acceleration. At the same time, Xilinx is also a leading chip supplier in the lidar market.
Xilinx and strategy analytics jointly launched the latest research on automotive lidar
·According to a latest report from strategy analytics supported by Xilinx, about US $1 billion has been invested in up to 100 different enterprises developing high-resolution lidar sensors. At the same time, the adoption of lidar sensors in the automotive market is also accelerating. Xilinx FPGA and adaptive SOC are being deployed in these systems because they can provide strong flexibility and flexibility, deliver high performance with low delay, and help achieve rapid product launch. During the Xilinx adapt conference, mark Fitzgerald, director of autonomous vehicle business of strategy analytics, will have a “stove talk” with Xilinx to discuss the insights in this report and make market forecasts. The time is scheduled at 11 a.m. Pacific time on September 15. In addition, Jim Kane, vice president of Baraja automotive engineering, will also attend the conversation between Xilinx and strategy analytics.
Xilinx data center platform and plug-in circuit board solutions continue to drive strong demand from large-scale cloud customers and financial technology market.
Azure synapse spark acceleration
·Microsoft announced that azure is now being adopted by Xilinx ® Alveo ™ Provide supported NP VM “FPGA as a service (FAAS)” infrastructure, which significantly improves the performance of Apache spark running on azure synapse analysis platform. Alveo acceleration can increase the data parsing speed of CSV and JSON file formats by up to 40 times, from 15MB / s with CPU core to 7.7gb/s with alveo card, thus eliminating a common bottleneck in spark performance. Since 90% of synapse spark workload consists of CSV / JSON / parquet files, azure has to provide significantly enhanced performance for its growing synapse customer base.
Real time video transcoding software development kit
·Xilinx also released the Xilinx video transcoding software development kit (video transcoding SDK), which enables streaming media to build new services or migrate existing services more easily and quickly than ever before, so as to give full play to the advantages of Xilinx acceleration. The video transcoding SDK is a complete solution, including from runtime driver to server resource management. It also supports ffmpeg and C + + / C + + API, so as to ensure that customers can start developing streaming media solutions immediately. The video transcoding SDK combined with Xilinx alveo u30 media accelerator card can provide industry-leading video transcoding performance. Compared with other alternative products, it has significantly improved channel density, lower power consumption and total cost of ownership.
Industry and medical
KRIA robot stack
·Xilinx is expanding the application scope of KRIA system module (SOM) beyond visual AI, and is providing robot experts with a complete set of functions and tools that are easy to use, adaptive and accelerated by hardware, called KRIA robot stack (KRS). KRS is built around the robot operating system (ROS, a framework for robot application development). It is composed of a set of integrated practical tools, which helps to improve the running speed of ROS 2 and gazebo and enable faster, safe and perfect robot behavior. KRS comes with an expanding set of reference designs. In the future, accelerated applications built on KRS and specifically for robots will be provided through Xilinx app store. KRS helps to use hardware acceleration on Xilinx devices and give full play to its advantages of high performance and low delay. Its approach is similar to Vitis AI, which helps AI developers implement machine learning inference models on Xilinx devices.
Vitis library for medical ultrasound
·Xilinx first added a library for ultra fast imaging to the extended Vitis unified software platform, which is optimized and oriented to the innovative versal AI engine. These libraries provide breakthrough performance for medical ultrasound applications. Using only a single versal device, a frame rate of more than 1000 frames per second can be achieved on a 64 channel ultrasonic instrument with 200 line resolution.
Wired and wireless communication
From the edge to the cloud, Xilinx FPGA and adaptive SOC continue to play a core role in promoting wired and wireless interconnected devices deployed in 5g around the world.
Xilinx zynq rfsoc DFE batch shipment
·Xilinx has begun to batch ship the new zynq rfsoc DFE to many radio customers around the world, including a top wireless system supplier. Zynq rfsoc DFE is a new type of radio platform. It integrates hardened digital front end (DFE) application specific modules used to provide 5g NR performance and reduce power consumption. At the same time, it also provides integrated programmable adaptive logic, which brings future oriented solutions for the continuously evolving 5g 3GPP standard and o-ran radio architecture. Zynq rfsoc DFE is gaining more and more momentum in the market because it not only optimizes energy efficiency, performance and cost, but also provides the flexibility and time to market advantages of programmable and adaptive SOC.
Figure: Xilinx zynq rfsoc DFE
On November 16, 2021, Xilinx adapt China virtual event will be launched soon.