一句话：论文设计并大规模部署了面向仓库级视频转码的专用加速器 VCU，实现 20–33 倍能效提升。
视频分享等负载占据互联网流量主体，摩尔定律放缓使得大规模视频处理需要专用硬件加速。本文介绍面向仓库级视频转码的新型加速器的设计与规模化部署，提出新的加速器构建块——视频编码单元（VCU），讨论数据中心规模下均衡系统的关键设计权衡，以及与大规模分布式软件系统的协同设计。作者在真实数据中心负载中评估该加速器，相较此前精心调优的非加速基线，能效提升 20–33 倍；设计还支持适应变化瓶颈、改进故障管理，并带来原系统无法实现的新工作负载能力。据称这是首个大规模仓库级环境中的视频加速工作。
Video sharing (e.g., YouTube, Vimeo, Facebook, TikTok) accounts for the majority of internet traffic, and video processing is also foundational to several other key workloads (video conferencing, virtual/augmented reality, cloud gaming, video in Internet-of-Things devices, etc.). The importance of these workloads motivates larger video processing infrastructures and – with the slowing of Moore’s law – specialized hardware accelerators to deliver more computing at higher efficiencies. This paper describes the design and deployment, at scale, of a new accelerator targeted at warehouse-scale video transcoding. We present our hardware design including a new accelerator building block – the video coding unit (VCU) – and discuss key design trade-offs for balanced systems at data center scale and co-designing accelerators with large-scale distributed software systems. We evaluate these accelerators “in the wild" serving live data center jobs, demonstrating 20-33x improved efficiency over our prior well-tuned non-accelerated baseline. Our design also enables effective adaptation to changing bottlenecks and improved failure management, and new workload capabilities not otherwise possible with prior systems. To the best of our knowledge, this is the first work to discuss video acceleration at scale in large warehouse-scale environments.
@inproceedings{paperbot6121,
title = {Warehouse-scale video acceleration: co-design and deployment in the wild},
author = {Parthasarathy Ranganathan and Daniel Stodolsky and Jeff Calow and Jeremy Dorfman and Marisabel Guevara and Clinton Wills Smullen IV and Aki Kuusela and Raghu Balasubramanian and Sandeep Bhatia and Prakash Chauhan and Anna Cheung and In Suk Chong and Niranjani Dasharathi and Jia Feng and Brian Fosco and Samuel Foss and Ben Gelb and Sara J. Gwin and Yoshiaki Hase and Da-ke He and C. Richard Ho and Roy W. Huffman Jr. and Elisha Indupalli and Indira Jayaram and Poonacha Kongetira and Cho Mon Kyaw and Aaron Laursen and Yuan Li and Fong Lou and Kyle A. Lucke and JP Maaninen and Ramon Macias and Maire Mahony and David Alexander Munday and Srikanth Muroor and Narayana Penukonda and Eric Perkins-Argueta and Devin Persaud and Alex Ramirez and Ville-Mikko Rautio and Yolanda Ripley and Amir Salek and Sathish Sekar and Sergey N. Sokolov and Rob Springer and Don Stark and Mercedes Tan and Mark S. Wachsler and Andrew C. Walton and David A. Wickeraad and Alvin Wijaya and Hon Kwan Wu},
booktitle = {ASPLOS},
year = {2021},
doi = {10.1145/3445814.3446723}
}