Accelerating Mobile Language Model via Speculative Decoding and NPU-Coordinated Execution

• URL: http://arxiv.org/abs/2510.15312v3
• Date: Thu, 23 Oct 2025 09:30:23 GMT
• Title: Accelerating Mobile Language Model via Speculative Decoding and NPU-Coordinated Execution
• Authors: Zhiyang Chen, Daliang Xu, Haiyang Shen, Mengwei Xu, Shangguang Wang, Yun Ma,
• Abstract summary: sd.npu is a framework that integrates speculative decoding with dynamic hardware scheduling to accelerate context-aware text generation on mobile devices.<n>Experiments show consistent improvements of up to 3.8x in generation speed and 4.7x in energy efficiency compared with existing mobile inference solutions.
• Score: 10.577037037457465
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Enhancing on-device large language models (LLMs) with contextual information from local data enables personalized and task-aware generation, powering use cases such as intelligent assistants and UI agents. While recent developments in neural processors have substantially improved the efficiency of prefill on mobile devices, the token-by-token generation process still suffers from high latency and limited hardware utilization due to its inherently memory-bound characteristics. This work presents sd.npu, a mobile inference framework that integrates speculative decoding with dynamic hardware scheduling to accelerate context-aware text generation on mobile devices. The framework introduces three synergistic components: (1) adaptive execution scheduling, which dynamically balances compute graphs between prefill and decoding phases; (2) context-aligned drafting, which improves speculative efficiency through lightweight online calibration to current tasks; and (3) hardware-efficient draft extension, which reuses and expands intermediate sequences to improve processing parallelism and reduce verification cost. Experiments on multiple smartphones and representative workloads show consistent improvements of up to 3.8x in generation speed and 4.7x in energy efficiency compared with existing mobile inference solutions. Component-level analysis further validates the contribution of each optimization.

Related papers

• IRG-MotionLLM: Interleaving Motion Generation, Assessment and Refinement for Text-to-Motion Generation [54.36300724708094]
  Assessment and refinement tasks act as crucial bridges to enable bidirectional knowledge flow between understanding and generation.<n>We introduce IRG-MotionLLM, the first model that seamlessly interleaves motion generation, assessment, and refinement to improve generation performance.
  arXiv  Detail & Related papers  (2025-12-11T15:16:06Z)
• MOBIUS: Big-to-Mobile Universal Instance Segmentation via Multi-modal Bottleneck Fusion and Calibrated Decoder Pruning [91.90342432541138]
  Scaling up model size and training data has advanced foundation models for instance-level perception.<n>High computational cost limits adoption on resource-constrained platforms.<n>We introduce a new benchmark for efficient segmentation on both high-performance computing platforms and mobile devices.
  arXiv  Detail & Related papers  (2025-10-16T18:00:00Z)
• CollaPipe: Adaptive Segment-Optimized Pipeline Parallelism for Collaborative LLM Training in Heterogeneous Edge Networks [57.95170323315603]
  We introduce CollaPipe, a distributed learning framework that integrates collaborative pipeline parallelism with federated aggregation to support self-evolving networks.<n>In CollaPipe, the encoder part is adaptively partitioned into variable-sized segments and deployed across mobile devices for pipeline-parallel training, while the decoder is deployed on edge servers to handle generative tasks.<n>To enhance training efficiency, we formulate a joint optimization problem that adaptively allocates model segments, micro-batches, bandwidth, and transmission power.
  arXiv  Detail & Related papers  (2025-09-24T07:54:01Z)
• POLARON: Precision-aware On-device Learning and Adaptive Runtime-cONfigurable AI acceleration [0.0]
  This work presents a SIMD-enabled, multi-precision MAC engine that performs efficient multiply-accumulate operations.<n>The architecture incorporates a layer adaptive precision strategy to align computational accuracy with workload sensitivity.<n>Results demonstrate up to 2x improvement in PDP and 3x reduction in resource usage compared to SoTA designs.
  arXiv  Detail & Related papers  (2025-06-10T13:33:02Z)
• Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention [32.48360534726024]
  We present NSA, a Natively trainable Sparse Attention mechanism that integrates algorithmic innovations with hardware-aligned optimizations.<n>NSA employs a dynamic hierarchical sparse strategy, combining coarse-grained token compression with fine-grained token selection to preserve both global context awareness and local precision.
  arXiv  Detail & Related papers  (2025-02-16T11:53:44Z)
• Characterizing Mobile SoC for Accelerating Heterogeneous LLM Inference [11.755602920122803]
  HeteroInfer is the fastest LLM inference engine in mobile devices which supports GPU-NPU heterogeneous execution.<n> Evaluation shows that HeteroInfer delivers a 1.34x to 6.02x end-to-end speedup over state-of-the-art GPU-NPU engines.
  arXiv  Detail & Related papers  (2025-01-11T02:42:02Z)
• Dovetail: A CPU/GPU Heterogeneous Speculative Decoding for LLM inference [31.901686946969786]
  Dovetail is an inference method that leverages the complementary characteristics of heterogeneous devices and the advantages of speculative decoding.<n>Dovetail achieves inference speedups ranging from 1.79x to 10.1x across different devices, while maintaining consistency and stability in the distribution of generated texts.
  arXiv  Detail & Related papers  (2024-12-25T15:45:18Z)
• Task-Oriented Real-time Visual Inference for IoVT Systems: A Co-design Framework of Neural Networks and Edge Deployment [61.20689382879937]
  Task-oriented edge computing addresses this by shifting data analysis to the edge. Existing methods struggle to balance high model performance with low resource consumption. We propose a novel co-design framework to optimize neural network architecture.
  arXiv  Detail & Related papers  (2024-10-29T19:02:54Z)
• MELTing point: Mobile Evaluation of Language Transformers [8.238355633015068]
  We explore the current state of mobile execution of Large Language Models (LLMs) We have created our own automation infrastructure, MELT, which supports the headless execution and benchmarking of LLMs on device. We evaluate popular instruction fine-tuned LLMs and leverage different frameworks to measure their end-to-end and granular performance.
  arXiv  Detail & Related papers  (2024-03-19T15:51:21Z)
• Energy-efficient Task Adaptation for NLP Edge Inference Leveraging Heterogeneous Memory Architectures [68.91874045918112]
  adapter-ALBERT is an efficient model optimization for maximal data reuse across different tasks. We demonstrate the advantage of mapping the model to a heterogeneous on-chip memory architecture by performing simulations on a validated NLP edge accelerator.
  arXiv  Detail & Related papers  (2023-03-25T14:40:59Z)
• Confident Adaptive Language Modeling [95.45272377648773]
  CALM is a framework for dynamically allocating different amounts of compute per input and generation timestep. We demonstrate the efficacy of our framework in reducing compute -- potential speedup of up to $times 3$ -- while provably maintaining high performance.
  arXiv  Detail & Related papers  (2022-07-14T17:00:19Z)
• DevFormer: A Symmetric Transformer for Context-Aware Device Placement [12.400790776196667]
  We present DevFormer, a transformer-based architecture for addressing the complex and computationally demanding problem of hardware design optimization. Our approach addresses this limitation by introducing strong inductive biases such as relative positional embeddings and action-permutation symmetricity. We show that DevFoemer outperforms state-of-the-art methods in both simulated hardware, leading to improved performances while reducing the number of components by more than $30$%.
  arXiv  Detail & Related papers  (2022-05-26T08:36:35Z)
• Multi-Exit Semantic Segmentation Networks [78.44441236864057]
  We propose a framework for converting state-of-the-art segmentation models to MESS networks. specially trained CNNs that employ parametrised early exits along their depth to save during inference on easier samples. We co-optimise the number, placement and architecture of the attached segmentation heads, along with the exit policy, to adapt to the device capabilities and application-specific requirements.
  arXiv  Detail & Related papers  (2021-06-07T11:37:03Z)
• Reconfigurable Intelligent Surface Assisted Mobile Edge Computing with Heterogeneous Learning Tasks [53.1636151439562]
  Mobile edge computing (MEC) provides a natural platform for AI applications. We present an infrastructure to perform machine learning tasks at an MEC with the assistance of a reconfigurable intelligent surface (RIS) Specifically, we minimize the learning error of all participating users by jointly optimizing transmit power of mobile users, beamforming vectors of the base station, and the phase-shift matrix of the RIS.
  arXiv  Detail & Related papers  (2020-12-25T07:08:50Z)

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.