Practical Resource Management in Power-Constrained, High Performance Computing

Power management is one of the key research challenges on the path to exascale. Supercomputers today are designed to be worst-case power provisioned, leading to two main prob-lems — limited application performance and under-utilization of procured power. In this paper, we propose RMAP, a practical, low-overhead resource manager targeted at future power-constrained clus-ters. The goals for RMAP are to improve application per-formance as well as system power utilization, and thus min-imize the average turnaround time for all jobs. Within RMAP, we design and analyze an adaptive policy, which derives job-level power bounds in a fair-share manner and supports overprovisioning and power-aware backfilling. Our results show that our new policy increases system power uti-lization while adhering to strict job-level power bounds and leads to 31.2 % (18.5 % on average) and 53.8 % (36.07 % on average) faster average turnaround time when compared to worst-case provisioning and naive overprovisioning respec-tively. 1