paper-with-me

홈 › Papers

TokenWeave: Efficient Compute-Communication Overlap for Distributed LLM Inference

2025-05-16 · Raja Gond, Nipun Kwatra, Ramachandran Ramjee

Distributed inference of large language models (LLMs) can introduce overheads of up to 20% even over GPUs connected via high-speed interconnects such as NVLINK. Multiple techniques have been proposed to mitigate these overheads by decomposing computations into finer-grained tasks and overlapping communication with sub-tasks as they complete. However, fine-grained decomposition of a large computation into many smaller computations on GPUs results in overheads. Further, the communication itself uses many streaming multiprocessors (SMs), adding to the overhead. We present TokenWeave to address these challenges. TokenWeave proposes a Token-Splitting technique that divides the tokens in the inference batch into two approximately equal subsets in a wave-aware manner. The computation of one subset is then overlapped with the communication of the other. In addition, TokenWeave optimizes the order of the layer normalization computation with respect to communication operations and implements a novel fused AllReduce-RMSNorm kernel carefully leveraging Multimem instruction support available on NVIDIA Hopper GPUs. These optimizations allow TokenWeave to perform communication and RMSNorm using only 2-8 SMs. Moreover, our kernel enables the memory bound RMSNorm to be overlapped with the other batch's computation, providing additional gains. Our evaluations demonstrate up to 29% latency gains and up to 26% throughput gains across multiple models and workloads. In several settings, TokenWeave results in better performance compared to an equivalent model with all communication removed.

📄 PDF Abstract BibTeX arXiv:2505.11329

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
RMSNorm 설명 없음

Similar Papers 제목 키워드 기반

CUCo: An Agentic Framework for Compute and Communication Co-design

2026-03-02 · Yoga Sri Varshan Varadharajan, Bodun Hu, Saurabh Agarwal, Aditya Akella arxiv

Computation and communication in distributed LLM training and inference are traditionally optimized in isolation; expert-crafted systems such as DeepEP, FLUX, and TokenWeave show the potential of co-design but require de…

T3: Transparent Tracking & Triggering for Fine-grained Overlap of Compute & Collectives

2024-01-30 · Suchita Pati, Shaizeen Aga, Mahzabeen Islam, Nuwan Jayasena 외

Large Language Models increasingly rely on distributed techniques for their training and inference. These techniques require communication across devices which can reduce scaling efficiency as the number of devices incre…

EnergyLens: Predictive Energy-Aware Exploration for Multi-GPU LLM Inference Optimization

2026-05-14 · Zhiye Song, Kyungmi Lee, Eun Kyung Lee, Xin Zhang 외 arxiv

We present EnergyLens, an end-to-end framework for energy-aware large language model (LLM) inference optimization. As LLMs scale, predicting and reducing their energy footprint has become critical for sustainability and …

CommFuse: Hiding Tail Latency via Communication Decomposition and Fusion for Distributed LLM Training

2026-04-27 · Rezaul Karim, Austin Wen, Wang Zongzuo, Weiwei Zhang 외 arxiv

The rapid growth in the size of large language models has necessitated the partitioning of computational workloads across accelerators such as GPUs, TPUs, and NPUs. However, these parallelization strategies incur substan…

Computational Efficiency

TicTac: Accelerating Distributed Deep Learning with Communication Scheduling

2018-03-08 · Sayed Hadi Hashemi, Sangeetha Abdu Jyothi, Roy H. Campbell

State-of-the-art deep learning systems rely on iterative distributed training to tackle the increasing complexity of models and input data. The iteration time in these communication-heavy systems depends on the computati…

Deep LearningScheduling