Ù�Ù�ضÙ�ع اÙ�بØ

Ù�Ù�ضÙ�ع اÙ�Ø¨Ø - International...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
International Journal of Computational Intelligence Research. ISSN 0973-1873 Vol.4, No.1 (2008), pp. 17–26 © Research India Publications http://www.ijcir.info A Survey of the State of the Art in Performance Modeling and Prediction of Parallel and Distributed Computing Systems Sabri Pllana, Ivona Brandic and Siegfried Benkner University of Vienna, Institute of Scientific Computing Nordbergstrasse 15, 1090 Vienna, Austria {pllana,brandic,sigi}@par.univie.ac.at Abstract: Performance is one of the key features of parallel and distributed computing systems. Therefore, in the past a significant research effort was invested in the development of approaches for performance modeling and prediction of parallel and distributed computing systems. In this paper we identify the trends, contributions, and drawbacks of the state of the art approaches. We describe a wide range of the performance modeling approaches that spans from the high- level mathematical modeling to the detailed instruction-level simulation. For each approach we describe how the program and machine are modeled and estimate the model development and evaluation effort, the efficiency, and the accuracy. Furthermore, we present an overall evaluation of the described approaches. 1. Introduction The solution of resource-demanding scientific and engineering computational problems involves the execution of programs on parallel and distributed computing machines, in order to solve large problems or to reduce the time to solution for a single problem [13]. However, the development of this kind of computing systems is an expensive and time-consuming endeavor. For instance, the development cost of the Earth Simulator Center (ESC) [10] was about US$350 million [43], and its development took about five years. While the life of parallel and distributed computing machines is commonly up to five years long, the life of parallel and distributed programs is up to 30 years [9]. Therefore, it is important to have means for the performance evaluation of programs not only on existing machines, but also on machines that are under development or being planned. Because the performance is a key indicator of computing systems, the performance evaluation was a preoccupation of many computer scientists in the past [1, 35, 18, 23, 25]. The commonly used techniques for the performance evaluation of computing systems include: measurement, mathematical modeling, and simulation. Each of these techniques has its limitations. Measurement techniques require that the system under study is available for experimentation, and their applicability is limited to only existing systems. Mathematical performance models usually lack the system’s structural information, and therefore, are not suitable for the model based performance analysis. The model-based performance analysis involves the modification of structure of the model to reflect system structural changes, in order to predict what would be the performance of the system under
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 10

Ù�Ù�ضÙ�ع اÙ�Ø¨Ø - International...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online