Title :
Effects of Low-Quality Computation Time Estimates in Policed Schedulers
Author :
Wozniak, Justin M. ; Jiang, Yingxin ; Striegel, Aaron
Author_Institution :
Notre Dame Univ., IN
Abstract :
Researchers conducting computer simulations can often provide estimates of computation time for a given type of simulation, which may be used by the compute cluster to aid in resource allocation and scheduling. However, the low quality of these estimates can cause deadline misses and unpredictable behavior. This problem is exacerbated on complex compute resources, clusters, and grids. Past-deadline jobs may be killed to provide resources for others, but the effect on the throughput of whole batches not fully understood. In this paper, we examine models for simple job schedulers and examine the quality of the deadline guarantee given. Simulation results based on actual runtimes are provided and discussed
Keywords :
digital simulation; grid computing; resource allocation; scheduling; computer simulations; deadline guarantee; job scheduling; low-quality computation time estimates; policed schedulers; resource allocation; Computational modeling; Computer simulation; Distributed computing; Grid computing; Processor scheduling; Real time systems; Resource management; Runtime; Scheduling algorithm; Throughput;
Conference_Titel :
Simulation Symposium, 2007. ANSS '07. 40th Annual
Conference_Location :
Norfolk, VA
Print_ISBN :
0-7695-2814-7
DOI :
10.1109/ANSS.2007.16