# PowerInfer-2: Fast Large Language Model Inference on a Smartphone

arXiv: https://arxiv.org/abs/2406.06282 · HTML: https://mlsanity.com/paper/2406.06282/

## Key facts

- Cited by 103 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 54 references that are in the map.
- First posted on arXiv 10 Jun 2024.

## Builds on (most cited first, 24 of 24)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads](https://mlsanity.com/paper/2401.10774/) (2024)
- [LLM-Pruner: On the Structural Pruning of Large Language Models](https://mlsanity.com/paper/2305.11627/) (2023)
- [EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty](https://mlsanity.com/paper/2401.15077/) (2024)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [QuIP: 2-Bit Quantization of Large Language Models With Guarantees](https://mlsanity.com/paper/2307.13304/) (2023)
- [Break the Sequential Dependency of LLM Inference Using Lookahead Decoding](https://mlsanity.com/paper/2402.02057/) (2024)
- [PowerInfer: Fast Large Language Model Serving with a Consumer-grade GPU](https://mlsanity.com/paper/2312.12456/) (2023)
- [RoleLLM: Benchmarking, Eliciting, and Enhancing Role-Playing Abilities of Large Language Models](https://mlsanity.com/paper/2310.00746/) (2023)
- [MoEfication: Transformer Feed-forward Layers are Mixtures of Experts](https://mlsanity.com/paper/2110.01786/) (2021)
- [The Lazy Neuron Phenomenon: On Emergence of Activation Sparsity in Transformers](https://mlsanity.com/paper/2210.06313/) (2022)
- [A Survey of Resource-efficient LLM and Multimodal Foundation Models](https://mlsanity.com/paper/2401.08092/) (2024)
- [ProSparse: Introducing and Enhancing Intrinsic Activation Sparsity within Large Language Models](https://mlsanity.com/paper/2402.13516/) (2024)
- [The Emergence of Essential Sparsity in Large Pre-trained Models: The Weights that Matter](https://mlsanity.com/paper/2306.03805/) (2023)
- [CATS: Contextually-Aware Thresholding for Sparsity in Large Language Models](https://mlsanity.com/paper/2404.08763/) (2024)
- [Transformer-Lite: High-efficiency Deployment of Large Language Models on Mobile Phone GPUs](https://mlsanity.com/paper/2403.20041/) (2024)
- [STI: Turbocharge NLP Inference at the Edge via Elastic Pipelining](https://mlsanity.com/paper/2207.05022/) (2022)

## Built on by (most cited first, 30 of 103)

- [Small Language Models are the Future of Agentic AI](https://mlsanity.com/paper/2506.02153/) (2025)
- [From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Models](https://mlsanity.com/paper/2406.16838/) (2024)
- [Small Language Models: Survey, Measurements, and Insights](https://mlsanity.com/paper/2409.15790/) (2024)
- [On-Device Language Models: A Comprehensive Review](https://mlsanity.com/paper/2409.00088/) (2024)
- [Fast On-device LLM Inference with NPUs](https://mlsanity.com/paper/2407.05858/) (2024)
- [HOBBIT: A Mixed Precision Expert Offloading System for Fast MoE Inference](https://mlsanity.com/paper/2411.01433/) (2024)
- [LServe: Efficient Long-sequence LLM Serving with Unified Sparse Attention](https://mlsanity.com/paper/2502.14866/) (2025)
- [Densing Law of LLMs](https://mlsanity.com/paper/2412.04315/) (2024)
- [Large Language Model Inference Acceleration: A Comprehensive Hardware Perspective](https://mlsanity.com/paper/2410.04466/) (2024)
- [BlueLM-V-3B: Algorithm and System Co-Design for Multimodal Large Language Models on Mobile Devices](https://mlsanity.com/paper/2411.10640/) (2024)
- [A Survey: Collaborative Hardware and Software Design in the Era of Large Language Models](https://mlsanity.com/paper/2410.07265/) (2024)
- [Configurable Foundation Models: Building LLMs from a Modular Perspective](https://mlsanity.com/paper/2409.02877/) (2024)
- [Sparsing Law: Towards Large Language Models with Greater Activation Sparsity](https://mlsanity.com/paper/2411.02335/) (2024)
- [Characterizing Mobile SoC for Accelerating Heterogeneous LLM Inference](https://mlsanity.com/paper/2501.14794/) (2025)
- [Bitnet.cpp: Efficient Edge Inference for Ternary LLMs](https://mlsanity.com/paper/2502.11880/) (2025)
- [TroL: Traversal of Layers for Large Language and Vision Models](https://mlsanity.com/paper/2406.12246/) (2024)
- [CoreInfer: Accelerating Large Language Model Inference with Semantics-Inspired Adaptive Sparse Activation](https://mlsanity.com/paper/2410.18311/) (2024)
- [IC-Cache: Efficient Large Language Model Serving via In-context Caching](https://mlsanity.com/paper/2501.12689/) (2025)
- [RetroInfer: A Vector Storage Engine for Scalable Long-Context LLM Inference](https://mlsanity.com/paper/2505.02922/) (2025)
- [MoQAE: Mixed-Precision Quantization for Long-Context LLM Inference via Mixture of Quantization-Aware Experts](https://mlsanity.com/paper/2506.07533/) (2025)
- [Efficient LLM Inference using Dynamic Input Pruning and Cache-Aware Masking](https://mlsanity.com/paper/2412.01380/) (2024)
- [Agent.xpu: Efficient Scheduling of Agentic LLM Workloads on Heterogeneous SoC](https://mlsanity.com/paper/2506.24045/) (2025)
- [Scaling LLM Test-Time Compute with Mobile NPU on Smartphones](https://mlsanity.com/paper/2509.23324/) (2025)
- [Understanding Large Language Models in Your Pockets: Performance Study on COTS Mobile Devices](https://mlsanity.com/paper/2410.03613/) (2024)
- [Cognitive Edge Computing: A Comprehensive Survey on Optimizing Large Models and AI Agents for Pervasive Deployment](https://mlsanity.com/paper/2501.03265/) (2025)
- [SpecEdge: Scalable Edge-Assisted Serving Framework for Interactive LLMs](https://mlsanity.com/paper/2505.17052/) (2025)
- [FloE: On-the-Fly MoE Inference on Memory-constrained GPU](https://mlsanity.com/paper/2505.05950/) (2025)
- [SparseLoRA: Accelerating LLM Fine-Tuning with Contextual Sparsity](https://mlsanity.com/paper/2506.16500/) (2025)
- [CoreMatching: A Co-adaptive Sparse Inference Framework with Token and Neuron Pruning for Comprehensive Acceleration of Vision-Language Models](https://mlsanity.com/paper/2505.19235/) (2025)
- [MNN-LLM: A Generic Inference Engine for Fast Large Language Model Deployment on Mobile Devices](https://mlsanity.com/paper/2506.10443/) (2025)

## Abstract (arXiv)

Large language models (LLMs) on smartphones enable real-time AI assistance and privacy-preserving, offline operation. However, resource constraints of smartphones limit current deployments to small language models (SLMs), significantly compromising their capabilities. This paper introduces PowerInfer-2, a smartphone-based framework that enables fast inference for LLMs exceeding the memory capacity. The key insight is decomposing matrix operations into neuron clusters as the basic processing unit, which enables flexible scheduling and efficient I/O-computation pipelining. PowerInfer-2 leverages this neuron-cluster-based design in both computation and storage. For computation, neuron clusters with dense activations are processed on NPU, while sparse clusters use CPU. The storage engine provides a fine-grained pipeline mechanism that coordinates cluster-level computation and I/O operations, enhanced by a segmented neuron cache to reduce I/O activities. PowerInfer-2 achieves up to a 27.8x speed increase compared to state-of-the-art frameworks. PowerInfer-2 is the first system to serve a 47B LLM on a smartphone, achieving 11.68 tokens/s. Notably, these performance improvements preserve model quality with negligible accuracy degradation.