A Workload Characterization Study of I/O Intensive Parallel Programs

Characterizing the I/O requirements of parallel applications that manipulate huge amounts of data, such as scientific codes, is critical to the achievement of good application performance and to the effective use of parallel systems. In this paper we formulate a compositional stochastic model of the behavior of I/O intensive scientific applications, which can be applied at various granularity levels of characterization. Based on the observation of the interaction of CPU and I/O activity in a set of scientific codes, we exercise the stochastic model. The model is in excellent agreement with experimental data in the set of the examined codes. For the model to be used for performance prediction, we also propose a set of functional forms for forecasting the scalability of computation and I/O components of an application. The complexity of the functional form adopted affects the accuracy of performance prediction. 1 Introduction In recent years, parallel scientific applications that manipu..