Energy Considerations of Large Language Model Inference and Efficiency Optimizations
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917997915930624 |
|---|---|
| author | Fernandez, Jared Na, Clara Tiwari, Vashisth Bisk, Yonatan Luccioni, Sasha Strubell, Emma |
| author_facet | Fernandez, Jared Na, Clara Tiwari, Vashisth Bisk, Yonatan Luccioni, Sasha Strubell, Emma |
| contents | As large language models (LLMs) scale in size and adoption, their computational and environmental costs continue to rise. Prior benchmarking efforts have primarily focused on latency reduction in idealized settings, often overlooking the diverse real-world inference workloads that shape energy use. In this work, we systematically analyze the energy implications of common inference efficiency optimizations across diverse Natural Language Processing (NLP) and generative Artificial Intelligence (AI) workloads, including conversational AI and code generation. We introduce a modeling approach that approximates real-world LLM workflows through a binning strategy for input-output token distributions and batch size variations. Our empirical analysis spans software frameworks, decoding strategies, GPU architectures, online and offline serving settings, and model parallelism configurations. We show that the effectiveness of inference optimizations is highly sensitive to workload geometry, software stack, and hardware accelerators, demonstrating that naive energy estimates based on FLOPs or theoretical GPU utilization significantly underestimate real-world energy consumption. Our findings reveal that the proper application of relevant inference efficiency optimizations can reduce total energy use by up to 73% from unoptimized baselines. These insights provide a foundation for sustainable LLM deployment and inform energy-efficient design strategies for future AI infrastructure. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_17674 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Energy Considerations of Large Language Model Inference and Efficiency Optimizations Fernandez, Jared Na, Clara Tiwari, Vashisth Bisk, Yonatan Luccioni, Sasha Strubell, Emma Computation and Language Machine Learning As large language models (LLMs) scale in size and adoption, their computational and environmental costs continue to rise. Prior benchmarking efforts have primarily focused on latency reduction in idealized settings, often overlooking the diverse real-world inference workloads that shape energy use. In this work, we systematically analyze the energy implications of common inference efficiency optimizations across diverse Natural Language Processing (NLP) and generative Artificial Intelligence (AI) workloads, including conversational AI and code generation. We introduce a modeling approach that approximates real-world LLM workflows through a binning strategy for input-output token distributions and batch size variations. Our empirical analysis spans software frameworks, decoding strategies, GPU architectures, online and offline serving settings, and model parallelism configurations. We show that the effectiveness of inference optimizations is highly sensitive to workload geometry, software stack, and hardware accelerators, demonstrating that naive energy estimates based on FLOPs or theoretical GPU utilization significantly underestimate real-world energy consumption. Our findings reveal that the proper application of relevant inference efficiency optimizations can reduce total energy use by up to 73% from unoptimized baselines. These insights provide a foundation for sustainable LLM deployment and inform energy-efficient design strategies for future AI infrastructure. |
| title | Energy Considerations of Large Language Model Inference and Efficiency Optimizations |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2504.17674 |