High Performance Biomedical Computing

高性能生物医学计算计划的目标是识别和解决那些可以受益于高性能计算和通信硬件、现代软件工程原理和高效算法的复杂生物医学问题。这项工作包括为美国国立卫生研究院提供高性能的并行计算机系统，以及为生物医学应用开发并行算法。使用高性能并行计算机，生物医学科学家可以极大地减少完成计算密集型任务所需的时间，并可以采用新的方法来处理实验数据。这可能允许在计算中包含更多的数据，确定更准确的结果，减少完成长时间计算所需的时间，或者实施新的算法或更现实的模型。高性能并行计算还允许生物医学科学家分析和研究大型数据集，这些数据集在传统的顺序或向量处理机上无法在实际的计算机时间内处理。通过高带宽网络连接和交互式用户界面，实验室或诊所的生物医学研究人员可以通过研究人员的计算机工作站轻松访问并行计算。为了应对这些计算挑战，计算生物科学与工程实验室(CBEL)正在为广泛的生物医学应用开发算法，在这些应用中，计算速度和高级可视化技术非常重要。这些领域包括基因芯片数据分析、电子显微图像处理、医学成像、电子顺磁共振成像和光谱分析、人类遗传连锁分析、通过核磁共振光谱学和X射线结晶学确定蛋白质结构以及分子动力学模拟。最终目标是提供高性能的并行计算、科学数据库以及科学数据分析和可视化工具，以促进NIH所做的科学工作。在开发这些计算要求很高的应用程序时，CBEL正在研究以下高性能计算问题：(1)将一个问题划分成许多部分，这些部分可以在不同的处理器上独立执行；(2)设计这些部分，使计算负载可以均匀地分布在可用的处理器上或动态平衡；(3)设计算法，使处理器的数量是一个参数，并且算法可以为可用的机器动态配置；(4)开发工具和环境来产生可移植的并行程序；(5)将交互数据分析和可视化工具结合到用户环境中；(6)监控系统性能；(7)证明给定机器上的并行算法满足其要求；(8)评估现代并行计算机体系结构在生物医学应用中的性能特征；(9)开发和提供高性能的数据库系统，用于通过Web接口对科学数据进行归档和分析。

The goals of the High Performance Biomedical Computing Program are to identify and solve those complex biomedical problems that can benefit from high performance computing and communications hardware, modern software engineering principles, and efficient algorithms. This effort includes providing high performance parallel computer systems for the NIH and development of parallel algorithms for biomedical applications. Using a high performance parallel computer, biomedical scientists can greatly reduce the time it takes to complete computationally intensive tasks and can adopt new approaches for processing experimental data. This may allow for the inclusion of more data in a calculation, the determination of a more accurate result, a reduction in the time needed to complete a long calculation, or the implementation of a new algorithm or more realistic model. High performance parallel computing also allows biomedical scientists to analyze and study large datasets that cannot be processed within a practical amount of computer time on conventional sequential or vector processing machines. With high bandwidth network connections and interactive user interfaces, parallel computing is readily accessible to a biomedical researcher in the laboratory or clinic at the investigator's computer workstation. In addressing these computational challenges, the Computational Bioscience and Engineering Laboratory (CBEL) is developing algorithms for a wide range of biomedical applications where computational speed and advanced visualization techniques are important. These include cDNA microarray data analysis, image processing of electron micrographs, medical imaging, electron paramagnetic resonance imaging and spectroscopy, human genetic linkage analysis, protein structure determination via nuclear magnetic resonance (NMR) spectroscopy and x-ray crystallography, and molecular dynamics simulations. The ultimate goal is to provide high performance parallel computing, scientific database, and sceintific data analysis and visualization tools to facilitate the science that is done at the NIH. While developing these computationally demanding applications, CBEL is investigating the following high performance computing issues: (1) partitioning a problem into many parts that can be independently executed on different processors; (2) designing the parts so that the computing load can be distributed evenly over the available processors or dynamically balanced; (3) designing algorithms so that the number of processors is a parameter and the algorithms can be configured dynamically for the available machine; (4) developing tools and environments for producing portable parallel programs; (5) incorporating interactive data analysis and visualization tools into the user environment; (6) monitoring system performance; (7) proving that a parallel algorithm on a given machine meets its specifications; (8) evaluating modern parallel computer architectures for their performance characteristics on biomedical applications; and (9) developing and providing high performance database systems for archive and analysis of scientific data via Web-interfaces.