Collaborative Research: CSR--AES: Interactive Parallel Platforms for Multi-Experiment Computational Studies

CSR-AES：多实验计算研究的交互并行平台科学家和工程师在从工程设计-空间探索到药物设计等领域越来越依赖于涉及多个实验(运行模拟软件)的计算研究。这个协作项目为一个系统奠定了基础，该系统促进了涉及大量实验的交互计算研究，允许用户基于部分结果连续监控和指导研究如何展开。这项工作特别针对涉及在紧耦合并行机上运行的中小型模拟问题的工程设计应用，强调(1)对整个研究的更高级别的用户控制；(2)在进行新的计算时重用先前实验的数据；(3)依靠应用程序和系统软件之间更紧密的耦合来动态地管理系统资源；(4)基于通用组件体系结构(CCA)的软件重用和更具渗透性的系统/解算器级别接口的标准化。该系统体系结构建立在SCIRun2框架之上，并包含用于将高级用户规范转换为所需实验的模块、用于将并行机资源分配到实验以优化研究进度的模块、以及用于估计模拟时间和资源成本的数值方法和算法。该工作的主要影响预计在工程设计中，具有更广泛的适用性，可用于其他类别的生物医学和几何建模应用。该项目还将具有重要的教育应用，通过为学生创造动手模拟探索体验，并通过培训对特定科学领域、模拟框架和数值求解器接口有深入了解的系统计算机科学家。

CSR-AES: Interactive Parallel Platforms for Multi-ExperimentComputational StudiesScientists and engineers increasingly rely upon computational studiesinvolving multiple experiments (runs of simulation software) in domainsranging from engineering design-space exploration to drug design.This collaborative project develops foundations for a system thatfacilitates interactive computational studies involving large numbers ofexperiments, permitting the user to continuously monitor and steer howthe study unfolds, based on partial results. The work specificallytargets engineering design applications involving small-to-medium sizesimulation problems running on tightly-coupled parallel machines, andemphasizes (1) higher-level user control of the overall study; (2) reuseof data from prior experiments in carrying forward new computations; (3)dynamic management of system resources by relying on a tighter couplingbetween application and system software; and (4) software reuse based oncommon component architecture (CCA) compliance and standardization of amore permeable system-/solver-level interface.The system architecture builds upon the SCIRun2 framework andincorporates modules for translating high-level user specifications intorequired experiments, for allocating parallel machine resources toexperiments to optimize study progress, and numerical methods andalgorithms for estimating simulation time and resource costs.The primary impact of the work is expected in engineering design, withbroader applicability to additional classes of biomedical and geometricmodeling applications. The project will also have important educationalapplications, by creating hands-on simulated exploration experiences forstudents, and by training system computer scientists with a deepunderstanding of specific scientific domains, simulation frameworks, andnumerical solver interfaces.