Strategies for Software Agnostic Power Adjustments with Machine Learning#
Johannes Gebert (HLRS)
Abstract#
Machine learning-based approaches allow the power consumption of computing nodes to be adjusted to meet computational needs while saving energy. Further advances are possible, for example, by implementing the approach in hardware and in a software-independent manner. This talk will provide an overview of such strategies for adapting power consumption, runtime, and energy consumption to meet the needs of high-performance computing centers.