Maximizing Throughput of Overprovisioned HPC Data Centers Under a Strict Power Budget

Abstract

Abstract-Building future generation supercomputers while constraining their power consumption is one of the biggest challenges faced by the HPC community. For example, US Department of Energy has set a goal of 20 MW for an exascale (10 18 flops) supercomputer. To realize this goal, a lot of research is being done to revolutionize hardware design to build power efficient computers and network interconnects. In this work, we propose a software-based online resource management system that leverages hardware facilitated capability to constrain the power consumption of each node in order to optimally allocate power and nodes to a job. Our scheme uses this hardware capability in conjunction with an adaptive runtime system that can dynamically change the resource configuration of a running job allowing our resource manager to re-optimize allocation decisions to running jobs as new jobs arrive, or a running job terminates.We also propose a performance modeling scheme that estimates the essential power characteristics of a job at any scale. The proposed online resource manager uses these performance characteristics for making scheduling and resource allocation decisions that maximize the job throughput of the supercomputer under a given power budget. We demonstrate the benefits of our approach by using a mix of jobs with different powerresponse characteristics. We show that with a power budget of 4.75 MW, we can obtain up to 5.2X improvement in job throughput when compared with the SLURM scheduling policy that is power-unaware. We corroborate our results with real experiments on a relatively small scale cluster, in which we obtain a 1.7X improvement.

Energy-efficient computing for HPC workloads on heterogeneous manycore chips

“…This is also called overprovisioning. In our previous work ( [41,42]), we have shown significant improvement in performance of a data center by using overprovisioning under a strict power budget. We have also shown the benefit of using integer linear programming methods for improving the performance of applications on chips with low voltage operation under a strict power budget [47].…”

“…Kadayif et al [25] use integer linear programming for determining the optimal number of cores that will be used in executing each nest in the code of array-intensive applications under energy and performance constraints. Power Aware Resource Manager, PARM, proposed by Sarood et al [41] uses Integer Linear Program (ILP) to schedule and determine the optimal allocation of power and compute nodes to jobs submitted to a data center. Venugopalan et al [49] propose the use of ILP for optimal task scheduling on multiprocessors.…”

Runtime-Guided Mitigation of Manufacturing Variability in Power-Constrained Multi-Socket NUMA Nodes

“…Etinski et al [14] propose an LP-based job scheduling policy; Sarood et al [30] use performance modeling to make job scheduling decisions in power constraint system to improve job throughput; and Ellsworth et al [13] discuss a dynamic job scheduling algorithm, which when running under a system-wide power limit, detects unused power and redistributes it to nodes that can make use of it.…”

Using an Adaptive HPC Runtime System to Reconfigure the Cache Hierarchy

“…This parallel management, in addition to applications' persistence, empowers the runtime system to provide other important features such as load balancing [19], fault tolerance [23], efficient parallel I/O [24], [25], and power management [26], [27]. Therefore, an adaptive RTS orchestrates a control system [28].…”