# Energy-Efficient GPU DVFS for Fine-Tuning of SLMs on Resource-constrained Embedded Devices

arXiv: https://arxiv.org/abs/2607.05933 · HTML: https://mlsanity.com/paper/2607.05933/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 19 references that are in the map.
- First posted on arXiv 7 Jul 2026.

## Builds on (most cited first, 11 of 11)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [A Review on Edge Large Language Models: Design, Execution, and Applications](https://mlsanity.com/paper/2410.11845/) (2024)
- [Zeus: Understanding and Optimizing GPU Energy Consumption of DNN Training](https://mlsanity.com/paper/2208.06102/) (2022)
- [Characterizing LLM Inference Energy-Performance Tradeoffs across Workloads and GPU Scaling](https://mlsanity.com/paper/2501.08219/) (2025)
- [Federated Fine-Tuning of LLMs on the Very Edge: The Good, the Bad, the Ugly](https://mlsanity.com/paper/2310.03150/) (2023)
- [DVFO: Learning-Based DVFS for Energy-Efficient Edge-Cloud Collaborative Inference](https://mlsanity.com/paper/2306.01811/) (2023)
- [Camel: Energy-Aware LLM Inference on Resource-Constrained Devices](https://mlsanity.com/paper/2508.09173/) (2025)
- [Reducing Compute Waste in LLMs through Kernel-Level DVFS](https://mlsanity.com/paper/2601.08539/) (2026)
- [Performance Characterization of Containerized DNN Training and Inference on Edge Accelerators](https://mlsanity.com/paper/2312.07220/) (2023)

## Built on by (most cited first, 1 of 1)

- [Fine-Tuning a 3B-Parameter LLM on a Smartphone: Characterizing Sustained Training](https://mlsanity.com/paper/2610.06325/) (2026)

## Abstract (arXiv)

Dynamic Voltage Frequency Scaling (DVFS) on resource-constrained embedded GPU platforms is essential for energy-efficient small language model (SLM) fine-tuning, as privacy- and personalization-driven adaptation increasingly requires local execution and involves repeated forward-backward optimization over many mini-batches, making it substantially more time- and energy-intensive than single-pass inference. To this end, 1) we first characterize the fine-tuning behavior of representative encoder-only SLMs of BERT variants, and autoregressive decoder-only SLMs of Pythia variants on GLUE benchmarks. In addition to the characterizations, 2) we propose a simple yet effective ML-based model selection that selects energy-optimal GPU DVFS settings on resource-constrained embedded platforms. Our results on NVIDIA Jetson AGX Orin demonstrate average 13.11% energy savings (up to 26.73%) over MAXN Mode 0, which has no explicit power cap.