mlsanityData to 7 Oct 2026

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

arXiv 2607.05933 ↗First posted 7 Jul 2026cs.PF cs.LGcited by 0

Not on the map. Why?

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.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 11 of its 19 references that are in the map. First posted on arXiv 7 Jul 2026.

Builds on

11 of 19 references.

References · 11

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  3. 2023Pythia: A Suite for Analyzing Large Language Models Across Training and Scalingcited by 1,734
  4. 2024A Review on Edge Large Language Models: Design, Execution, and Applicationscited by 98
  5. 2022Zeus: Understanding and Optimizing GPU Energy Consumption of DNN Trainingcited by 78
  6. 2025Characterizing LLM Inference Energy-Performance Tradeoffs across Workloads and GPU Scalingcited by 19
  7. 2023Federated Fine-Tuning of LLMs on the Very Edge: The Good, the Bad, the Uglycited by 17
  8. 2025Camel: Energy-Aware LLM Inference on Resource-Constrained Devicescited by 14
  9. 2023DVFO: Learning-Based DVFS for Energy-Efficient Edge-Cloud Collaborative Inferencecited by 13
  10. 2026Reducing Compute Waste in LLMs through Kernel-Level DVFScited by 2
  11. 2023Performance Characterization of Containerized DNN Training and Inference on Edge Acceleratorscited by 1

Built on by

Not cited yet.