NGS: Collaborative Research: Reusable, Observation-based Performance Prediction across Platforms

本项目研究了一种基于观察的执行时间估计方法，用于网格环境下应用程序和资源调度的资源规划和使用估计。更具体地说，建议的方法将收集/管理/利用应用程序特征和性能结果，并在不连续的应用程序和硬件平台之间平等地传输此类信息。使用这些方法，一个应用程序在一个平台上执行的性能数据有助于预测另一个应用程序在另一个平台上的性能。这项研究的预期结果是一个元预测器，一个有效、高效和足够准确的跨平台性能预测工具，可以作为一项通用服务提供性能预测，以帮助网格用户进行长期研究规划和日常作业执行。这些方法将在生产平台上进行验证和评估，这些平台的应用程序代表国家相关的高端应用程序，如国家实验室生产代码。

This project, investigate an observation-based execution time estimation, approach for resource planning and usage estimation in the grid environment for application and resource scheduling. More specifically, the proposed approaches will collect/manage/utilize application characteristics and performance results, and equally transfer such information across disjoint applications and hardware platforms. With these approaches, performance data from one application's executions on one platform helps predict the performance of another application on another platform. The expected outcome of this research to be a meta-predictor, an effective, efficient and sufficiently accurate cross-platform performance prediction tool that can provide performance predictions as a general service to assist grid users in both their long-term research planning and their everyday job execution. These approaches will be validated and evaluated on production platforms with applications representative for nationally relevant high-end applications, such as National Lab production codes.