A Value-Oriented Job Scheduling Approach for Power-Constrained and Oversubscribed HPC Systems
- 16 January 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 31 (6) , 1419-1433
Abstract
In this article, we investigate limitations in the traditional value-based algorithms for a power-constrained HPC system and evaluate their impact on HPC productivity. We expose the trade-off between allocating system-wide power budget uniformly and greedily under different system-wide power constraints in an oversubscribed system. We experimentally demonstrate that, under the tightest power constraint, the mean productivity of the greedy allocation is 38 percent higher than the uniform allocation whereas, under the intermediate power constraint, the uniform allocation has a mean productivity of 6 percent higher than the greedy allocation. We then propose a new algorithm that adapts its behavior to deliver the combined benefits of the two allocation strategies. We design a methodology with online retraining capability to create application-specific power-execution time models for a class of HPC applications. These models are used in predicting the execution time of an application on the available resources at the time of making scheduling decisions in the power-aware algorithms. We evaluate the proposed algorithm using emulation and simulation environments, and show that our adaptive strategy results in improving HPC resource utilization while delivering a mean productivity that is almost the same as the best performing algorithm across various system-wide power constraints.Keywords
Funding Information
- National Science Foundation (NSF CNS-1624668, CCF-1302693)
- Lawrence Livermore National Laboratory (DE-AC52-07NA27344 (LLNL-JRNL-775437))
This publication has 31 references indexed in Scilit:
- Practical Resource Management in Power-Constrained, High Performance Computing2015
- Electrical Grid and Supercomputing Centers: An Investigative Analysis of Emerging Opportunities and ChallengesInformatik Spektrum, 2014
- Experience with using the Parallel Workloads ArchiveJournal of Parallel and Distributed Computing, 2014
- Utility Functions and Resource Management in an Oversubscribed Heterogeneous Computing EnvironmentIEEE Transactions on Computers, 2014
- Productivity in High-Performance Computing2008
- Productivity Metrics and Models for High Performance ComputingThe International Journal of High Performance Computing Applications, 2004
- High Performance Computing Productivity Model SynthesisThe International Journal of High Performance Computing Applications, 2004
- Measuring High Performance Computing ProductivityThe International Journal of High Performance Computing Applications, 2004
- A Framework for Measuring Supercomputer ProductivityThe International Journal of High Performance Computing Applications, 2004
- A scheduling algorithm for tasks described by Time Value FunctionReal-Time Systems, 1996