Swing: Short-cutting Rings for Higher Bandwidth Allreduce
CoRR(2024)
摘要
The allreduce collective operation accounts for a significant fraction of the
runtime of workloads running on distributed systems. One factor determining its
performance is the distance between communicating nodes, especially on networks
like torus, where a higher distance implies multiple messages being forwarded
on the same link, thus reducing the allreduce bandwidth. Torus networks are
widely used on systems optimized for machine learning workloads (e.g., Google
TPUs and Amazon Trainium devices), as well as on some of the Top500
supercomputers. To improve allreduce performance on torus networks we introduce
Swing, a new algorithm that keeps a low distance between communicating nodes by
swinging between torus directions. Our analysis and experimental evaluation
show that Swing outperforms by up to 3x existing allreduce algorithms for
vectors ranging from 32B to 128MiB, on different types of torus and torus-like
topologies, regardless of their shape and size.
更多查看译文
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