清华-赛灵思FPGA机器学习研讨会

课程内容

针对机器学习数据分析和视频流处理的Xilinx可重构加速平台

摘要：当前GPU已经无法继续满足机器学习、数据分析和视频处理计算密集应用的要求。微软、亚马逊等主要云服务提供商开始部署FPGA作为未来海量数据应用平台。FPGA的可编程能力和高性能计算能力特别适合于高速变化的市场。本报告将介绍Xilinx针对机器学习等大数据应用的硬件、软件最新进展。

Amit Gupta is Senior Director of Software Engineering at Xilinx responsible for new product introduction. He is also leading Xilinx investment in startups and academic research program to vastly deploy FPGA as acceleration platform. Amit’s team also works cross-functionally across organizations at Xilinx enabling new methodologies for productivity and quality improvement. 

reVision -针对嵌入式应用的Xilinx可重构平台

摘要： Xilinx reVision嵌入式可重构平台提供以下关键能力：

1）更快的响应时间；

2）针对最新算法和传感器的定制；

3）基于OpenCV、OpenCL、C/C++和Caffe等工具的软件编程能力

Ashish Sirasao is Distinguished Engineer in Xilinx Software and IP organization. He and his team are currently involved in acceleration activities for data-center with a focus on defining and implementing methodologies and hardware architecture for high-performance accelerators in the area of Machine Learning, Data Analytics, Genomics and Computer Vision using SDx acceleration platform. His team is also involved in enabling well known open-source software stacks in these areas so that customers can seamlessly deploy FPGA-based acceleration in data center environment.

日程安排