Performance Reproduction and Prediction of Selected Dynamic Loop Scheduling Experiments

摘要

Scientific applications are complex, large, and often exhibit irregular and stochastic behavior. The use of efficient loop scheduling techniques, from static to fully dynamic, in computationally-intensive applications is crucial for improving their performance, often degraded by load imbalance, on highperformance computing (HPC) platforms. A number of dynamic loop scheduling (DLS) techniques have been proposed between the late 1980's and early 2000's, and efficiently used in scientific applications. In most cases, the computing systems on which they have been tested and validated are no longer available. This work is concerned with the minimization of the sources of uncertainty in the implementation of DLS techniques to avoid unnecessary influences on the performance of scientific applications. Therefore, it is important to ensure that the DLS techniques employed in scientific applications today adhere to their original design goals and specifications. The goal of this work is to attain and increase the trust in the implementation of DLS techniques in today's studies. To achieve this goal, the performance of a selection of scheduling experiments from the 1992 original work that introduced factoring is reproduced and predicted via both, simulative and native experimentation. The scientific challenge is the reproduction of the performance of the past experiments with incomplete information, such as the computing system characteristics and the implementation details. The experiments show that the simulation reproduces the performance achieved on the past computing platform and accurately predicts the performance achieved on the present computing platform. The performance reproduction and prediction confirm that the present implementation of the DLS techniques considered both, in simulation and natively, adheres to their original description. The results confirm the hypothesis that reproducing experiments of identical scheduling scenarios on past and modern hardware leads to an entirely different behavior from expected.