The Design Process for Google's Training Chips: TPUv2 and TPUv3

IEEE Micro(2021)

引用 92|浏览487
暂无评分
摘要
Five years ago, few would have predicted that a software company like Google would build its own computers. Nevertheless, Google has been deploying computers for machine learning (ML) training since 2017, powering key Google services. These Tensor Processing Units (TPUs) are composed of chips, systems, and software, all co-designed in-house. In this paper, we detail the circumstances that led to this outcome, the challenges and opportunities observed, the approach taken for the chips, a quick review of performance, and finally a retrospective on the results. A companion paper describes the supercomputers built from these chips, the compiler, and a detailed performance analysis [Jou20].
更多
查看译文
关键词
design process,Google's training chips,TPUv2,TPUv3,machine learning training,ML,key Google services,tensor processing units,supercomputers
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要