arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2312.02493cs.DCcs.AIcs.CV
收藏

不可预测网络下实现最优分布式学习的灵活通信机制

Flexible Communication for Optimal Distributed Learning over Unpredictable Networks

Indiana University Bloomington(印第安纳大学伯明顿分校)

Sahil Tyagi, Martin Swany

AI总结:

针对分布式深度学习中通信成本与精度、效率的权衡问题,提出兼容Allreduce的带宽最优Topk压缩器及AG/AR灵活切换策略,通过多目标优化动态调压缩比,在不可预测网络下实现高精度与训练加速的平衡。

摘要:

梯度压缩通过发送更少的数值及其对应索引来缓解分布式深度学习中高昂的通信开销,通常采用Allgather(AG,全收集)操作实现。使用高压缩比(CR)训练可取得与DenseSGD相当的高精度,但因通信成本高导致并行扩展性较差(即并行效率低)。采用更低的压缩比可通过降低同步成本提升并行效率,但也会降低模型精度(统计效率)。此外,不同模型和压缩比带来的加速效果还会随网络延迟、有效带宽以及用于聚合的集合通信操作类型而变化。在很多场景下,对于相同数据量的交换,Allreduce(AR,全规约)这类集合操作的成本低于AG。本文提出了一种兼容AR的Topk压缩器,该压缩器具备带宽最优性,因此在特定网络配置下表现优于AG。我们开发了一种灵活通信策略,可根据当前设置下哪种集合操作更优在AG和AR之间切换,并将并行效率与统计效率之间的帕累托关系建模为多目标优化(MOO)问题,以动态调整压缩比、加速训练,同时仍能收敛到高精度。

英文摘要:

Gradient compression alleviates expensive communication in distributed deep learning by sending fewer values and its corresponding indices, typically via Allgather (AG). Training with high compression ratio (CR) achieves high accuracy like DenseSGD, but has lower parallel scaling due to high communication cost (i.e., parallel efficiency). Using lower CRs improves parallel efficiency by lowering synchronization cost, but degrades model accuracy as well (statistical efficiency). Further, speedup attained with different models and CRs also varies with network latency, effective bandwidth and collective op used for aggregation. In many cases, collectives like Allreduce (AR) have lower cost than AG to exchange the same amount of data. In this paper, we propose an AR-compatible Topk compressor that is bandwidth-optimal and thus performs better than AG in certain network configurations. We develop a flexible communication strategy that switches between AG and AR based on which collective is optimal in the current settings, and model the pareto-relationship between parallel and statistical efficiency as a multi-objective optimization (MOO) problem to dynamically adjust CR and accelerate training while still converging to high accuracy.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑