# Dissecting the Impact of Mobile DVFS Governors on LLM Inference Performance and Energy Efficiency

arXiv: https://arxiv.org/abs/2507.02135 · HTML: https://mlsanity.com/paper/2507.02135/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 43 references that are in the map.
- First posted on arXiv 2 Jul 2025.

## Builds on (most cited first, 13 of 13)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [TinyLlama: An Open-Source Small Language Model](https://mlsanity.com/paper/2401.02385/) (2024)
- [Splitwise: Efficient generative LLM inference using phase splitting](https://mlsanity.com/paper/2311.18677/) (2023)
- [MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases](https://mlsanity.com/paper/2402.14905/) (2024)
- [PowerInfer-2: Fast Large Language Model Inference on a Smartphone](https://mlsanity.com/paper/2406.06282/) (2024)
- [MNN: A Universal and Efficient Inference Engine](https://mlsanity.com/paper/2002.12418/) (2020)
- [LLMCad: Fast and Scalable On-device Large Language Model Inference](https://mlsanity.com/paper/2309.04255/) (2023)
- [MobileQuant: Mobile-friendly Quantization for On-device Language Models](https://mlsanity.com/paper/2408.13933/) (2024)
- [DVFO: Learning-Based DVFS for Energy-Efficient Edge-Cloud Collaborative Inference](https://mlsanity.com/paper/2306.01811/) (2023)
- [CLBlast: A Tuned OpenCL BLAS Library](https://mlsanity.com/paper/1705.05249/) (2017)
- [GPUReplay: A 50-KB GPU Stack for Client ML](https://mlsanity.com/paper/2105.05085/) (2021)

## Built on by (most cited first, 5 of 5)

- [LLM Inference at the Edge: Mobile, NPU, and GPU Performance Efficiency Trade-offs Under Sustained Load](https://mlsanity.com/paper/2603.23640/) (2026)
- [Parameter-Free Encoders Remain Viable for RDB Foundation Models](https://mlsanity.com/paper/2607.05476/) (2026)
- [SparseDVFS: Sparse-Aware DVFS for Energy-Efficient Edge Inference](https://mlsanity.com/paper/2603.21908/) (2026)
- [Mitigating Compiler Fusion-Induced Power Bursts in Mobile NPU Inference as the Battery Depletes](https://mlsanity.com/paper/2607.16555/) (2026)
- [Fine-Tuning a 3B-Parameter LLM on a Smartphone: Characterizing Sustained Training](https://mlsanity.com/paper/2610.06325/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) are increasingly being integrated into various applications and services running on billions of mobile devices. However, deploying LLMs on resource-limited mobile devices faces a significant challenge due to their high demand for computation, memory, and ultimately energy. While current LLM frameworks for mobile use three power-hungry components-CPU, GPU, and Memory-even when running primarily-GPU LLM models, optimized DVFS governors for CPU, GPU, and memory featured in modern mobile devices operate independently and are oblivious of each other. Motivated by the above observation, in this work, we first measure the energy-efficiency of a SOTA LLM framework consisting of various LLM models on mobile phones which showed the triplet mobile governors result in up to 40.4% longer prefilling and decoding latency compared to optimal combinations of CPU, GPU, and memory frequencies with the same energy consumption for sampled prefill and decode lengths. Second, we conduct an in-depth measurement study to uncover how the intricate interplay (or lack of) among the mobile governors cause the above inefficiency in LLM inference. Finally, based on these insights, we design FUSE - a unified energy-aware governor for optimizing the energy efficiency of LLM inference on mobile devices. Our evaluation using a ShareGPT dataset shows FUSE reduces the time-to-first-token and time-per-output-token latencies by 7.0%-16.9% and 25.4%-36.8% on average with the same energy-per-token for various mobile LLM models.