# EnergyLens: Predictive Energy-Aware Exploration for Multi-GPU LLM Inference Optimization

arXiv: https://arxiv.org/abs/2605.14249 · HTML: https://mlsanity.com/paper/2605.14249/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Efficient MoE training and inference systems, one of 12 directions in Mixture-of-experts.
- Builds on 18 of its 26 references that are in the map.
- First posted on arXiv 14 May 2026.

## Placed in

- Mixture-of-experts / [Efficient MoE training and inference systems](https://mlsanity.com/moe/efficient-moe-training-and-inference-systems/)

## Builds on (most cited first, 18 of 18)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving](https://mlsanity.com/paper/2401.09670/) (2024)
- [Quantifying the Carbon Emissions of Machine Learning](https://mlsanity.com/paper/1910.09700/) (2019)
- [DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scale](https://mlsanity.com/paper/2207.00032/) (2022)
- [Estimating the Carbon Footprint of BLOOM, a 176B Parameter Language Model](https://mlsanity.com/paper/2211.02001/) (2022)
- [Towards the Systematic Reporting of the Energy and Carbon Footprints of Machine Learning](https://mlsanity.com/paper/2002.05651/) (2020)
- [DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models](https://mlsanity.com/paper/2309.14509/) (2023)
- [Vidur: A Large-Scale Simulation Framework For LLM Inference](https://mlsanity.com/paper/2405.05465/) (2024)
- [ASTRA-sim2.0: Modeling Hierarchical Networks and Disaggregated Systems for Large-model Training at Scale](https://mlsanity.com/paper/2303.14006/) (2023)
- [LLMServingSim: A HW/SW Co-Simulation Infrastructure for LLM Inference Serving at Scale](https://mlsanity.com/paper/2408.05499/) (2024)
- [LLMCO2: Advancing Accurate Carbon Footprint Prediction for LLM Inferences](https://mlsanity.com/paper/2410.02950/) (2024)
- [Forecasting GPU Performance for Deep Learning Training and Inference](https://mlsanity.com/paper/2407.13853/) (2024)
- [Efficient and Adaptable Overlapping for Computation and Communication via Signaling and Reordering](https://mlsanity.com/paper/2504.19519/) (2025)
- [LLMServingSim 2.0: A Unified Simulator for Heterogeneous and Disaggregated LLM Serving Infrastructure](https://mlsanity.com/paper/2602.23036/) (2026)
- [Lumos: Efficient Performance Modeling and Estimation for Large-scale LLM Training](https://mlsanity.com/paper/2504.09307/) (2025)
- [Characterizing Compute-Communication Overlap in GPU-Accelerated Distributed Deep Learning: Performance and Power Implications](https://mlsanity.com/paper/2507.03114/) (2025)
- [EnergAIzer: Fast and Accurate GPU Power Estimation Framework for AI Workloads](https://mlsanity.com/paper/2604.20105/) (2026)

## Abstract (arXiv)

We present EnergyLens, an end-to-end framework for energy-aware large language model (LLM) inference optimization. As LLMs scale, predicting and reducing their energy footprint has become critical for sustainability and datacenter operations, yet existing approaches either require production-level code and expensive profiling or fail to accurately capture multi-GPU energy behavior. As a result, practitioners lack tools for deciding which optimizations to prioritize and for selecting among existing deployment configurations when exhaustive profiling is impractical. EnergyLens addresses this gap with an intuitive einsum-based interface that captures LLM specifications including fusion, parallelism, and compute-communication overlap, combined with load-imbalance-aware MoE modeling and an empirically driven communication energy model for multi-GPU settings. We validate EnergyLens on Llama3 and Qwen3-MoE across tensor-parallel and expert-parallel configurations, achieving mean absolute percentage errors (MAPEs) between 9.25% and 13.19% for multi-GPU prefill and decode energy, and 12.97% across SM allocations for Megatron-style overlap. Our energy-driven exploration reveals up to 1.47x and 52.9x energy variation across configurations in prefill and decode efficiency and motivates distributed serving. We further show that compute-communication overlap is difficult to optimize with intuition alone, but EnergyLens correctly identifies Pareto-optimal overlap configurations.