dorsal/arxiv
View SchemaReducing Compute Waste in LLMs through Kernel-Level DVFS
| Authors | Jeffrey Spaan, Kuan-Hsun Chen, Ana-Lucia Varbanescu |
|---|---|
| Categories | |
| ArXiv ID | 2601.08539vv1 |
| URL | https://arxiv.org/abs/2601.08539 |
| License | http://creativecommons.org/licenses/by/4.0/ |
Abstract
The rapid growth of AI has fueled the expansion of accelerator- or GPU-based data centers. However, the rising operational energy consumption has emerged as a critical bottleneck and a major sustainability concern. Dynamic Voltage and Frequency Scaling (DVFS) is a well-known technique used to reduce energy consumption, and thus improve energy-efficiency, since it requires little effort and works with existing hardware. Reducing the energy consumption of training and inference of Large Language Models (LLMs) through DVFS or power capping is feasible: related work has shown energy savings can be significant, but at the cost of significant slowdowns. In this work, we focus on reducing waste in LLM operations: i.e., reducing energy consumption without losing performance. We propose a fine-grained, kernel-level, DVFS approach that explores new frequency configurations, and prove these save more energy than previous, pass- or iteration-level solutions. For example, for a GPT-3 training run, a pass-level approach could reduce energy consumption by 2% (without losing performance), while our kernel-level approach saves as much as 14.6% (with a 0.6% slowdown). We further investigate the effect of data and tensor parallelism, and show our discovered clock frequencies translate well for both. We conclude that kernel-level DVFS is a suitable technique to reduce waste in LLM operations, providing significant energy savings with negligible slow-down.
{
"annotation_id": "b72c8ef4-62eb-4676-b531-7f4f4a8b145a",
"date_created": "2026-02-17T05:53:16.122000Z",
"date_modified": "2026-02-17T05:53:16.122000Z",
"file_hash": "3b1ca97954284c30d7c4e000a746b30b4902ec444195ff3316a71a9889aa71dc",
"private": false,
"record": {
"abstract": "The rapid growth of AI has fueled the expansion of accelerator- or GPU-based data centers. However, the rising operational energy consumption has emerged as a critical bottleneck and a major sustainability concern. Dynamic Voltage and Frequency Scaling (DVFS) is a well-known technique used to reduce energy consumption, and thus improve energy-efficiency, since it requires little effort and works with existing hardware. Reducing the energy consumption of training and inference of Large Language Models (LLMs) through DVFS or power capping is feasible: related work has shown energy savings can be significant, but at the cost of significant slowdowns. In this work, we focus on reducing waste in LLM operations: i.e., reducing energy consumption without losing performance. We propose a fine-grained, kernel-level, DVFS approach that explores new frequency configurations, and prove these save more energy than previous, pass- or iteration-level solutions. For example, for a GPT-3 training run, a pass-level approach could reduce energy consumption by 2% (without losing performance), while our kernel-level approach saves as much as 14.6% (with a 0.6% slowdown). We further investigate the effect of data and tensor parallelism, and show our discovered clock frequencies translate well for both. We conclude that kernel-level DVFS is a suitable technique to reduce waste in LLM operations, providing significant energy savings with negligible slow-down.",
"arxiv_id": "2601.08539",
"authors": [
"Jeffrey Spaan",
"Kuan-Hsun Chen",
"Ana-Lucia Varbanescu"
],
"categories": [
"cs.PF",
"cs.LG"
],
"license": "http://creativecommons.org/licenses/by/4.0/",
"title": "Reducing Compute Waste in LLMs through Kernel-Level DVFS",
"url": "https://arxiv.org/abs/2601.08539",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "d8922473-c7c5-42bc-884f-57f37ecdea0c",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}