Power-Capping Metric Evaluation for Improving Energy Efficiency in HPC Applications

TL;DR

This paper evaluates runtime power-capping strategies on exascale HPC systems, demonstrating how integrated CPU-GPU power management can enhance energy efficiency for compute-bound applications.

Contribution

It introduces a novel evaluation framework for power-capping metrics and demonstrates their effectiveness on NVIDIA GH200 architecture for exascale applications.

Findings

Power-capping can reduce energy consumption in GPU tasks.

Integrated CPU-GPU power management improves energy efficiency.

Modest energy savings have significant impact at exascale.

Abstract

With high-performance computing systems now running at exascale, optimizing power-scaling management and resource utilization has become more critical than ever. This paper explores runtime power-capping optimizations that leverage integrated CPU-GPU power management on architectures like the NVIDIA GH200 superchip. We evaluate energy-performance metrics that account for simultaneous CPU and GPU power-capping effects by using two complementary approaches: speedup-energy-delay and a Euclidean distance-based multi-objective optimization method. By targeting a mostly compute-bound exascale science application, the Locally Self-Consistent Multiple Scattering (LSMS), we explore challenging scenarios to identify potential opportunities for energy savings in exascale applications, and we recognize that even modest reductions in energy consumption can have significant overall impacts. Our results highlight how GPU task-specific dynamic power-cap adjustments combined with integrated CPU-GPU power steering can improve the energy utilization of certain GPU tasks, thereby laying the groundwork for future adaptive optimization strategies.