This article was originally published on December 21, 2018.
As this year comes to an end, we're taking a moment to reflect on the Adaptive Advantage blogs posted in 2018, covering product news, boards, IP, webinars, partner stories, etc. Here are the top 5 product news stories of 2018 selected by the editors.
Silicon product news dominated the headlines for much of the year. On the other end of the spectrum, boards and software news stories were popular enough to crack the top 5 as well. Read on to see the prominent product news stories of the year.
5. Machine Learning Is Everywhere and We’re Just Living It
With the DeePhi acquisition, Xilinx provides industry-leading capabilities in machine learning, specializing in deep compression technology. Xilinx now provides high-level tools to compile ML framework trained networks directly to optimized FPGA or SoC implementations.
For Zynq®-7000 SoCs and Zynq UltraScale+™ MPSoCs, the Deep Neural Network Development Kit (DNNDK) is designed as an integrated framework, which aims to simplify and accelerate deep learning applications’ development and deployment on DeePhi Deep Learning Processing Unit (DPU) platform. If you want to improve your deep learning performance, Download and try the DNNDK today on common development kits like ZCU102, ZCU104, and Ultra96! You will soon realize that this is a better workshop than Santa’s.
For Alveo™ accelerator cards and cloud providers, the Xilinx ML Suite provides support for many ML frameworks such as Caffe, MxNet, and Tensorflow as well as Python and RESTful APIs. The Xilinx ML Suite enables developers to optimize and deploy accelerated ML inference, especially low latency inferences. There are multiple ways to get started: AWS, Nimbix, or Alveo accelerator cards. If you want to try the Xilinx ML Suite, here is On-Demand tutorial on Amazon EC2 F1 FPGA instances.
4. Zynq UltraScale+ RFSoC ‒ An Unlikely Match Proves Successful
Zynq® UltraScale+™ RFSoC brings analog, digital, and embedded software design onto a single monolithic device for system robustness and provides another level of flexibility and adaptability. Zynq UltraScale+ RFSoCs integrate RF data converters and SD-FEC cores to one device. All Zynq UltraScale+ RFSoCs are in production. Also, Xilinx and partners have made ZCU111 evaluation kit, ZCU1275 Characterization Kit, and RF Data Converter Evaluation tool available for fast development and evaluation.
3. To Keep Or Not To Keep. PAM4-Capable Virtex UltraScale+ FPGAs Have Solutions For You
The industry is gearing up in anticipation of the change with new optics and standards. 58G PAM4 will be the next-generation interconnect and PAM4-capable Virtex® UltraScale+ FPGAs are here to enable not only next-generation optics and scalability but also to double performance while preserving existing legacy 25G backplanes. In 2018 at various tradeshows, Xilinx has shown our 58G PAM4 solution as well as the industry’s first 112G PAM4 demo.
Here are a couple of videos we demonstrated this year showing our technology leadership.
2. Alveo ‒ World’s Fastest Data Center and AI Accelerator Cards
Introduced at XDF in October 2018, Alveo™ Data Center accelerator cards deliver up to 20X real-time inference performance vs. high-end CPUs and up to 4X vs. high-end GPUs. Also revealed was the AMD EPYC and Xilinx Alveo accelerator cards, a world-record producing 30,000 image-per-second inference throughput server.
Alveo accelerator cards enable developers to adapt and optimize their algorithms for changing acceleration workloads, all with the same piece of hardware. For applications needing the highest performance but also the adaptability to accelerate any workload, test drive Alveo accelerator cards today.
1. Versal ACAP ‒ Ushering in the Era of Heterogeneous Compute for Any Application and Any Developer
The number 1 product news of 2018 at Xilinx was Versal® ACAP! At XDF 2018, Xilinx unveiled the Versal portfolio - the industry's first adaptive compute acceleration platform (ACAP). The Versal portfolio is the first platform to combine software programmability with domain-specific hardware adaptability. Rather than one fixed engine, the Versal ACAP draws on the strengths of CPUs, GPUs, and FPGAs to accelerate any application. Most importantly, the platform’s hardware and software can be programmed, optimized, and deployed by developers with any background (hardware or software). What are you waiting for? Check out our white papers, data sheet overview, AI Engine blog, and a video to get started. Don’t forget to sign up to stay informed as well!
In 2018, Xilinx brought significant progress in various technological innovations from FPGAs to accelerator cards to a new device category. Please share your ideas about the top 5 product news at Xilinx.
Happy Holidays and see you in 2019!
