Presentation

DescriptionThe two major driving forces that are leading centers to investigate dynamic power and energy management strategies are (1) limitations to the availability of resources from the electricity service provider, and (2) the desire to spend limited budgets on computational cycles rather than infrastructure requirements such as electricity. In addition, supercomputer systems have increasingly rapid, unpredictable, and large power fluctuations. In addition, electricity service providers may request supercomputing centers to change their timing and/or magnitude of demand to help address electricity supply constraints. To adapt to this new landscape, centers may employ energy and power-aware job scheduling and resource management (EPA-JSRM) strategies to dynamically, and in real-time, control their electricity demand. This poster summarizes the lessons learned from one of the first global surveys of supercomputing centers that are actively investigating such strategies.

Authors