High performance cluster need to consume lots of power. High power consumption is an important factor that restricts the development of system in both technical and economic ways. The power control technology has become a hot issue in the area of high performance computing systems and data center management. For different programs, reducing power consumption can exert different influence on performance. In order to minimize the losses of performance, adjustments of power consumption should be made based on the characteristics of programs. This paper aims at solving the problem of power control of nodes in high performance computing system. PCPI (Power Control Programming Interface) is designed to acquire real-time power consumption state and to provide function of dynamic power consumption control, so as to make effective power adjustments according to the operating characteristics of programs. In this way, power consumption can be reduced while the losses of performance are minimized.
[1]
Laxmikant V. Kalé,et al.
Optimizing power allocation to CPU and memory subsystems in overprovisioned HPC systems
,
2013,
2013 IEEE International Conference on Cluster Computing (CLUSTER).
[2]
Dong Li,et al.
PowerPack: Energy Profiling and Analysis of High-Performance Systems and Applications
,
2010,
IEEE Transactions on Parallel and Distributed Systems.
[3]
Laxmikant V. Kalé,et al.
Maximizing Throughput of Overprovisioned HPC Data Centers Under a Strict Power Budget
,
2014,
SC14: International Conference for High Performance Computing, Networking, Storage and Analysis.
[4]
Jean-Marc Menaud,et al.
Performance and Power Management for Cloud Infrastructures
,
2010,
2010 IEEE 3rd International Conference on Cloud Computing.
[5]
Martin Schulz,et al.
A Run-Time System for Power-Constrained HPC Applications
,
2015,
ISC.