Pith. sign in

REVIEW 12 cited by

LoongTrain: Efficient Training of Long-Sequence LLMs with Head-Context Parallelism

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2406.18485 v1 pith:3O3DC37R submitted 2024-06-26 cs.DC

LoongTrain: Efficient Training of Long-Sequence LLMs with Head-Context Parallelism

classification cs.DC
keywords loongtraintrainingllmsparallelismscalabilityefficiencyefficientlylong
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Efficiently training LLMs with long sequences is important yet challenged by the massive computation and memory requirements. Sequence parallelism has been proposed to tackle these problems, but existing methods suffer from scalability or efficiency issues. We propose LoongTrain, a novel system to efficiently train LLMs with long sequences at scale. The core of LoongTrain is the 2D-Attention mechanism, which combines both head-parallel and context-parallel techniques to break the scalability constraints while maintaining efficiency. We introduce Double-Ring-Attention and analyze the performance of device placement strategies to further speed up training. We implement LoongTrain with the hybrid ZeRO and Selective Checkpoint++ techniques. Experiment results show that LoongTrain outperforms state-of-the-art baselines, i.e., DeepSpeed-Ulysses and Megatron Context Parallelism, in both end-to-end training speed and scalability, and improves Model FLOPs Utilization (MFU) by up to 2.88x.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 12 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. HCMS: Head-Chunked Multi-Stream Pipeline for Communication-Computation Overlap in Long-Sequence Parallel Attention

    cs.DC 2026-07 unverdicted novelty 7.0

    HCMS partitions multi-head attention into chunks and pipelines them across dual CUDA streams to overlap communication and computation, delivering 10-17.5% speedup over Ulysses for 31K-56K token sequences.

  2. LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generation

    cs.CV 2026-05 unverdicted novelty 7.0

    LongLive-2.0 delivers an NVFP4 parallel infrastructure that enables direct training of long multi-shot autoregressive diffusion video models and achieves up to 2.15x training and 1.84x inference speedups on Blackwell ...

  3. Syncopate: Efficient Multi-GPU AI Kernels via Automatic Chunk-Centric Compute-Communication Overlap

    cs.DC 2026-01 conditional novelty 7.0

    A Triton-based compiler automatically generates fused distributed kernels with chunk-level compute-communication overlap, reporting 1.3x average and up to 4.7x speedups on multi-GPU workloads.

  4. Libra: Taming Attention Workload Skew in Long-Context LLM Training with Bounded Sequence Pool

    cs.DC 2026-07 conditional novelty 6.0

    Libra bounds attention load-balancing to fixed-size sequence pools and combines variance-reduced sequence placement with tiled attention pooling, achieving up to 2.54x end-to-end throughput over Ulysses on 256K- and 1...

  5. FlashCP: Load-Balanced Communication-Efficient Context Parallelism for LLM Training

    cs.DC 2026-06 unverdicted novelty 6.0

    FlashCP introduces Whole-Doc sharding, sharding-aware KV communication, and a heuristic for mixed sharding plans, claiming up to 1.63x speedup over prior CP methods for LLM training.

  6. Addressing Variable Heterogeneity in Distributed Multimodal Training with Entrain

    cs.DC 2026-05 unverdicted novelty 6.0

    Entrain reduces microbatch workload variability by up to 10.6x and improves multimodal LLM training throughput by 1.4x via static model parallelism and deferred hierarchical microbatch assignment.

  7. HexiSeq: Accommodating Long Context Training of LLMs over Heterogeneous Hardware

    cs.DC 2026-05 unverdicted novelty 6.0

    HexiSeq optimizes sequence and head partitioning across mixed GPUs to improve long-context LLM training throughput by up to 1.72x in simulations.

  8. CoCoDiff: Optimizing Collective Communications for Distributed Diffusion Transformer Inference Under Ulysses Sequence Parallelism

    cs.DC 2026-04 unverdicted novelty 6.0

    CoCoDiff achieves 3.6x average and 8.4x peak speedup for distributed DiT inference on up to 96 GPU tiles via tile-aware all-to-all, V-first scheduling, and selective V communication.

  9. DASH: Deterministic Attention Scheduling for High-throughput Reproducible LLM Training

    cs.LG 2026-01 conditional novelty 6.0

    DASH speeds up deterministic attention backprop by reversed and shift-based tile schedules, giving up to 1.28× throughput and ~5% end-to-end gains while keeping bitwise reproducibility.

  10. Syncopate: Efficient Multi-GPU AI Kernels via Automatic Chunk-Centric Compute-Communication Overlap

    cs.DC 2026-01 unverdicted novelty 6.0

    Syncopate automatically overlaps compute and communication at fine chunk granularity inside a single fused Triton kernel, yielding 1.3x average and up to 4.7x end-to-end speedup on multi-GPU workloads.

  11. MTraining: Distributed Dynamic Sparse Attention for Efficient Ultra-Long Context Training

    cs.CL 2025-10 conditional novelty 6.0

    MTraining scales LLM training to 512K-token contexts on 32 A100 GPUs by integrating dynamic sparse training patterns with balanced and hierarchical sparse ring attention, achieving up to 6x throughput gains without ac...

  12. MAGI-1: Autoregressive Video Generation at Scale

    cs.CV 2025-05 unverdicted novelty 6.0

    MAGI-1 is a 24B-parameter autoregressive video world model that predicts denoised frame chunks sequentially with increasing noise to enable causal, scalable, streaming generation up to 4M token contexts.