Major chip companies are jumping into data center accelerator game. This includes Xilinx, which seeks to mainstream FPGAs in the datacenter and dive into the hyper-scale race.
Xilinx Latest News & Update: Meeting Demands Of The Market
Xilinx is a rapidly growing company that already has seven customers with world's largest data centers. These include Alibaba, Amazon, Baidu, Facebook, Google, Microsoft and Tancent. These technologies turn a specific algorithm at least 10 times faster in order to meet the demand in applying a machine learning, ultra high definition streaming video and complex data analytics.
EETimes says that Xilinx launches its FPGA reconfigurable acceleration chips to take the mainstream fast. It features a library, framework integrations, developer boards and OpenStack support. This is in response to what hyper-scale companies' demand for a reconfigurable acceleration that can handle machine learning, transcoding to data analytics, networking and storage. Hence, Xilinx took the lead into the growing market by getting designed into Microsoft's project hyper-scale acceleration fabric.
More Xilinx Features
According to Forbes, Xilinx focuses on the workloads of large and fast growing footprints in hyper-scale data centers. Machine learning in the form of Artificial Intelligence has become the new demand while SQL analytics has become pervasive across hyper-scale applications. On the other hand, the live 4k streaming becomes big business around the globe. Hence, Xilinx targets the 3 fastest growing workloads with the reprogrammable acceleration chips. These are the machine learning, video transcoding for live 4k streaming and the SQL queries for data analytics.
The company will enable silicon optimization performance workloads. This can be done by swapping the most optimal design bit stream. Therefore, when comparing with x86 server CPUs, Xilinx's FPGAs will give the fastest way 40x better efficiency. Xilinx will also boost the company's offer with up to 6 times efficiency over the FPGAs.
.
