EnergyLens: Predictive Energy-Aware Exploration for Multi-GPU LLM Inference Optimization
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910218824187904 |
|---|---|
| author | Song, Zhiye Lee, Kyungmi Lee, Eun Kyung Zhang, Xin Eilam, Tamar Chandrakasan, Anantha P. |
| author_facet | Song, Zhiye Lee, Kyungmi Lee, Eun Kyung Zhang, Xin Eilam, Tamar Chandrakasan, Anantha P. |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_14249 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | EnergyLens: Predictive Energy-Aware Exploration for Multi-GPU LLM Inference Optimization Song, Zhiye Lee, Kyungmi Lee, Eun Kyung Zhang, Xin Eilam, Tamar Chandrakasan, Anantha P. Machine Learning 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. |
| title | EnergyLens: Predictive Energy-Aware Exploration for Multi-GPU LLM Inference Optimization |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2605.14249 |