Session 3: Optimizing Energy with Performance in Mind

Speaker

Abstract

Discover how to minimize energy consumption while maintaining performance targets using advanced optimization frameworks. This session introduces recent advances in energy optimization for training and inference.

Resources

• Slides (PDF)

References & Helpful Links

• DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency (HPCA '25)
• The ML.ENERGY Benchmark: Toward Automated Inference Energy Measurement and Optimization (NeurIPS '25 D&B)
• Zeus: Understanding and Optimizing GPU Energy Consumption of DNN Training (NSDI '23)
• Reducing Energy Bloat in Large Model Training (SOSP '24)