A domain-specific supercomputer for training deep neural networks
A domain-specific supercomputer for training deep neural networks
N. Jouppi,D. Yoon,5 Authors,David A. Patterson
2020 · DOI: 10.1145/3360307
Communications of the ACM · 293 Citations
TLDR
Google's TPU supercomputers train deep neural networks 50x faster than general-purpose supercomputing running a high-performance computing benchmark.
Abstract
Google's TPU supercomputers train deep neural networks 50x faster than general-purpose supercomputers running a high-performance computing benchmark.