Abstract
Abstract The need for greater performance in high performance computing systems combined with rising costs of electricity to power these systems motivates the need for energy-efficient resource management. Driven by the requirements of the Extreme Scale Systems Center at Oak Ridge National Laboratory, we address the problem of scheduling dynamically-arriving tasks to machines in an oversubscribed and energy-constrained heterogeneous distributed computing environment. Our goal is to maximize total "utility" earned by the system, where the utility of a task is defined by a monotonically-decreasing function that represents the value of completing that task at different times. To address this problem, we design four energy-aware resource allocation heuristics and compare their performance to heuristics from the literature. For our given energy-constrained environment, we also design an energy filtering technique that helps some heuristics regulate their energy consumption by allowing tasks to only consume up to an estimated fair-share of energy. Extensive sensitivity analyses of the heuristics in environments with different levels of heterogeneity show that heuristics with the ability to balance both energy consumption and utility exhibit the best performance because they save energy for use by future tasks.
| Original language | English |
|---|---|
| Article number | 100 |
| Pages (from-to) | 14-30 |
| Number of pages | 17 |
| Journal | Sustainable Computing: Informatics and Systems |
| Volume | 5 |
| DOIs | |
| State | Published - Mar 1 2015 |
Funding
This research used resources of the National Center for Computational Sciences at Oak Ridge National Laboratory, supported by the Extreme Scale Systems Center at ORNL, which is supported by the Department of Defense. Additional support was provided by a National Science Foundation Graduate Research Fellowship, and by NSF Grants CCF-1302693 and CCF-1252500 . Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. This research also used the CSU ISTeC Cray System supported by NSF Grant CNS-0923386 . The authors thank Mark Oxley and Eric Jonardi for their valuable comments. A preliminary version of portions of this material appeared in [34] . This work builds upon the workshop paper with the design of three new heuristics (i.e., the weighted heuristics) for the energy-constrained environment that perform better than their corresponding conference counterparts (i.e., the non-weighted heuristics). We analyze why the weighting technique outperforms the filtering technique even though they both regulate the energy consumption and hit the energy constraint close to the end of the day. We also design a method to create ETCs of lower and higher Task Machine Affinity than the example environment, i.e., different degrees of heterogeneity. We perform extensive heuristic parameter tuning tests and analyze the performance of all the heuristics in the example, low, and high TMA environments.
| Funders | Funder number |
|---|---|
| CSU ISTeC Cray System | CNS-0923386 |
| National Science Foundation | CCF-1252500, CCF-1302693 |
| U.S. Department of Defense |
Keywords
- Energy-aware resource management
- Energy-constrained computing
- Heterogeneous distributed computing
- High performance computing system
