ShadowNPU: System and Algorithm Co-design for NPU-Centric On-Device LLM Inference
Lists
Only you see this
Abstract
On-device running Large Language Models (LLMs) is nowadays a critical enabler towards preserving user privacy. We observe that the attention operator falls back from the special-purpose NPU to the general-purpose CPU/GPU because of quantization sensitivity in state-of-the-art frameworks. This fallback results in a degraded user experience and increased complexity in system scheduling. To this end, this paper presents shadowAttn, a system-algorithm codesigned sparse attention module with minimal reliance on CPU/GPU by only sparsely calculating the attention on a tiny portion of tokens. The key idea is to hide the overhead of estimating the important tokens with a NPU-based pilot compute. Further, shadowAttn proposes insightful techniques such as NPU compute graph bucketing, head-wise NPU-CPU/GPU pipeline and per-head fine-grained sparsity ratio to achieve high accuracy and efficiency. shadowAttn delivers the best performance with highly limited CPU/GPU resource; it requires much less CPU/GPU resource to deliver on-par performance of SoTA frameworks.
BibTeX
@misc{yin2025shadownpu,
author = {Wangsong Yin and Daliang Xu and Mengwei Xu and Gang Huang and Xuanzhe Liu},
title = {{ShadowNPU: System and Algorithm Co-design for NPU-Centric On-Device LLM Inference}},
year = {2025},
eprint = {2508.16703},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2508.16703},
}Loading related content...