AMDP: Asynchronous Multi-Directional Pipeline Parallelism for Large-Scale Models Training

• URL: http://arxiv.org/abs/2605.29664v1
• Date: Thu, 28 May 2026 09:25:12 GMT
• Title: AMDP: Asynchronous Multi-Directional Pipeline Parallelism for Large-Scale Models Training
• Abstract summary: AMDP limits the first stage of each pipeline to process at most two minibatches before backpropagation.<n>Experiments on GPT- and BERT-style models demonstrate that AMDP significantly accelerates training while preserving convergence.
• Score: 8.956450951670979
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Pipeline parallelism is essential for large-scale model training, but existing asynchronous approaches often degrade convergence due to parameter mismatch between forward and backward passes. We propose Asynchronous Multi-Directional Pipeline parallelism (AMDP) to mitigate this issue while sustaining high utilization. AMDP limits the first stage of each pipeline to process at most two minibatches before backpropagation, bounding the number of parameter updates between forward and backward passes. To alleviate the resulting pipeline bubbles, AMDP launches multiple concurrent pipelines and adapts their number according to pipeline depth. In addition, AMDP accumulates gradients across minibatches and applies them in a single update, ensuring that only a bounded number of minibatches experience parameter mismatch, limited to within one optimization step. Experiments on GPT- and BERT-style models demonstrate that AMDP significantly accelerates training while preserving convergence.

Related papers

• One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining [9.465438386458823]
  A Pipeline Parallelism eliminates pipeline bubbles, maximizing throughput at the cost of staleness.<n>We show that optimizing under staleness depends strongly on choice rather than being an intrinsic limitation.<n>Our strategies bridge the performance gap with synchronous training, highlighting the practical potential of asynchronous pipeline parallelism at scale.
  arXiv  Detail & Related papers  (2026-06-29T17:57:50Z)
• Breaking the Bubble: Asynchronous Pipeline Parallel Training with Bounded Weight Inconsistency [5.093144724934585]
  PACI is a bubble-free pipeline method that bounds forward/backward version drift without weight stashing, prediction, additional parameter copies, or global synchronization.<n>In GPT-style language-model pretraining, PACI matches the stability and final perplexity of synchronous 1F1B-flush, retains the same peak memory footprint, and achieves fully utilized pipeline throughput.<n>These results show that forward/backward inconsistency need not be eliminated: when explicitly bounded, it can be safely traded for substantial efficiency gains.
  arXiv  Detail & Related papers  (2026-06-05T22:33:57Z)
• PARM: Pipeline-Adapted Reward Model [60.769414637325326]
  Reward models (RMs) are central to aligning large language models (LLMs) with human preferences, powering advanced decoding strategies.<n>While most prior work focuses on single-step generation, real-world applications increasingly adopt multi-stage pipelines.<n>We investigate this through code generation for optimization, constructing a pipeline that integrates reward models into both formulation and solution stages.
  arXiv  Detail & Related papers  (2026-04-20T14:29:08Z)
• TimelyFreeze: Adaptive Parameter Freezing Mechanism for Pipeline Parallelism [9.595296799577499]
  We propose TimelyFreeze, which models the pipeline schedule as a directed acyclic graph and solves a linear program to compute optimal freeze ratios.<n> Experiments show that TimelyFreeze achieves up to 40% training throughput improvement on LLaMA-8B with comparable accuracy.
  arXiv  Detail & Related papers  (2026-02-05T15:24:11Z)
• AsyncMesh: Fully Asynchronous Optimization for Data and Pipeline Parallelism [54.8494905524997]
  We introduce asynchronous updates across both parallelism axes, relaxing the co-location requirement.<n>We provide convergence guarantees for both sparse averaging and asynchronous updates.<n>Experiments on large-scale language models demonstrate that our approach matches the performance of the fully synchronous baseline.
  arXiv  Detail & Related papers  (2026-01-30T01:24:47Z)
• AdaPtis: Reducing Pipeline Bubbles with Adaptive Pipeline Parallelism on Heterogeneous Models [59.7059443712562]
  AdaPtis is a training system for large language models (LLMs) that supports adaptive pipeline parallelism.<n>Extensive experiments show that AdaPtis achieves an average speedup of 1.42x (up to 2.14x) over Megatron-LM I-1F1B.
  arXiv  Detail & Related papers  (2025-09-28T08:05:13Z)
• HelixPipe: Efficient Distributed Training of Long Sequence Transformers with Attention Parallel Pipeline Parallelism [14.067070576474086]
  As transformer sequence lengths grow, existing pipeline parallelisms incur suboptimal performance due to the quadratic attention computation and the substantial memory overhead.<n>We propose HelixPipe, a novel pipeline parallelism for long sequence transformer training.<n>It introduces attention parallel partition, which schedules attention computations of different micro batches across different pipeline stages in parallel, reducing pipeline bubbles.<n>It employs a two-fold first-in-last-out micro batch schedule to balance memory usage and overlap communication with fragmentation.
  arXiv  Detail & Related papers  (2025-07-01T03:11:18Z)
• Nesterov Method for Asynchronous Pipeline Parallel Optimization [59.79227116582264]
  We introduce a variant of Nesterov Accelerated Gradient (NAG) for asynchronous optimization in Pipeline Parallelism.<n>Specifically, we modify the look-ahead step in NAG to effectively address the staleness in gradients.<n>We theoretically prove that our approach converges at a sublinear rate in the presence of fixed delay in gradients.
  arXiv  Detail & Related papers  (2025-05-02T08:23:29Z)
• Self-Refining Diffusion Samplers: Enabling Parallelization via Parareal Iterations [53.180374639531145]
  Self-Refining Diffusion Samplers (SRDS) retain sample quality and can improve latency at the cost of additional parallel compute.<n>We take inspiration from the Parareal algorithm, a popular numerical method for parallel-in-time integration of differential equations.
  arXiv  Detail & Related papers  (2024-12-11T11:08:09Z)
• BitPipe: Bidirectional Interleaved Pipeline Parallelism for Accelerating Large Models Training [5.7294516069851475]
  BitPipe is a bidirectional interleaved pipeline parallelism for accelerating large models training. We show that BitPipe improves the training throughput of GPT-style and BERT-style models by 1.05x-1.28x compared to the state-of-the-art synchronous approaches.
  arXiv  Detail & Related papers  (2024-10-25T08:08:51Z)
• Stochastic Optimization with Laggard Data Pipelines [65.20044914532221]
  We show that "dataechoed" extensions of common optimization methods exhibit provable improvements over their synchronous counterparts. Specifically, we show that in convex optimization with minibatches, data echoing affords speedups on the curvature-dominated part of the convergence rate, while maintaining the optimal statistical rate.
  arXiv  Detail & Related papers  (2020-10-26T14:55:31Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.