English

OptPipe: Memory- and Scheduling-Optimized Pipeline Parallelism for LLM Training

Distributed, Parallel, and Cluster Computing 2025-10-08 v1 Artificial Intelligence Optimization and Control

Abstract

Pipeline parallelism (PP) has become a standard technique for scaling large language model (LLM) training across multiple devices. However, despite recent progress in reducing memory consumption through activation offloading, existing approaches remain largely heuristic and coarse-grained, often overlooking the fine-grained trade-offs between memory, computation, and scheduling latency. In this work, we revisit the pipeline scheduling problem from a principled optimization perspective. We observe that prevailing strategies either rely on static rules or aggressively offload activations without fully leveraging the interaction between memory constraints and scheduling efficiency. To address this, we formulate scheduling as a constrained optimization problem that jointly accounts for memory capacity, activation reuse, and pipeline bubble minimization. Solving this model yields fine-grained schedules that reduce pipeline bubbles while adhering to strict memory budgets. Our approach complements existing offloading techniques: whereas prior approaches trade memory for time in a fixed pattern, we dynamically optimize the tradeoff with respect to model structure and hardware configuration. Experimental results demonstrate that our method consistently improves both throughput and memory utilization. In particular, we reduce idle pipeline time by up to 50% under the same per-device memory limit, and in some cases, enable the training of larger models within limited memory budgets.

Keywords

Cite

@article{arxiv.2510.05186,
  title  = {OptPipe: Memory- and Scheduling-Optimized Pipeline Parallelism for LLM Training},
  author = {Hongpei Li and Han Zhang and Huikang Liu and Dongdong Ge and Yinyu Ye},
  journal= {arXiv preprint arXiv:2510.05186},
  year   = {2025}
}

Comments

Use Mathematical Programming to model Pipeline Parallelism with Offloading to balance efficiency and memory requirement

R2 v1 2026-07-01T06:19:49.266Z