CAREER: Fast Methods for Particle Dynamics and Their Applications

该项目将开发创新技术和软件，用于n体求解及其应用的所谓“快速方法”。n体解算器计算一组相互作用粒子的行为；例子从太阳系（9个行星通过引力相互作用）到计算机芯片（数十亿个原子吸引或排斥电子）。这些求解器的“快速”方法使用诸如树之类的数据结构来最小化计算次数。通常，数据结构总结了一组粒子的相互作用，消除了计算所有成对相互作用的需要。管理数据结构使这些方法难以编程，特别是在现代并行机器上。该项目解决了算法问题，串行和并行性能，可移植性，以及在选定应用中快速粒子动力学方法的I/O要求。具体而言，该项目在以下领域做出了研究贡献：*误差控制和边界条件：该项目将扩展最近在变度多极中的工作，以改善非均匀粒子分布的误差边界，同时最大限度地减少增加的计算。*强大的并行算法和库：该项目将建立校准的便携式快速求解程序库，能够适应动态计算平台，如多处理器集群。*数据分析和I/O控制：该项目将研究能够有效处理粒子动力学模拟产生的大量数据的压缩方案。教育计划包括各级的各种举措。K-12科学教育中的计算机应用（CAKE）项目将向大学预科学生介绍高性能计算和网络的使用，使用基于研究图书馆的gui作为教学工具。一系列关于并行计算及其应用的课程、研讨会和讲习班将把普渡大学的教育扩展到本科生和研究生水平。

This project will develop innovative techniques and software for so-called"fast methods" for n-body solvers and their applications. An n-body solvercomputes the behavior of a set of interacting particles; examples run fromthe solar system (9 planets interacting through gravity) to computer chips(billions of atoms that attract or repel electrons). "Fast" methods forthese solvers use data structures such as trees to minimize the number ofcomputations. Typically, the data structure summarizes the interactions ofa group of particles, eliminating the need to compute all of the pairwiseinteractions. Managing the data structure makes these methods difficult toprogram, particularly on modern parallel machines.This project addresses algorithmic issues, serial and parallel performance,portability, and I/O requirements for fast methods for particle dynamics inselected applications. Specifically, the project makes researchcontributions in the following areas:* Error Control and Boundary Conditions: This project will extend recentwork in variable degree multipoles to improve error bounds in non-uniformparticle distributions while minimizing the added computation.* Robust Parallel Algorithms and Libraries: This project will buildcalibrated portable libraries of fast solvers capable of adapting todynamic computational platforms like clusters of multiprocessors.* Data Analysis and I/O Control: This project will investigate compressionschemes capable of effectively handling the huge amounts of data generatedby particle dynamics simulations.The education plan includes a variety of initiatives at various levels. TheComputer Applications in K-12 Education in Science (CAKE) program willintroduce pre-college students to the use of high performance computing andnetworking, using GUIs based on the research libraries as teaching tools.A series of courses, seminars, and workshops on parallel computing and itsapplications will extend the education to the undergraduate and graduatelevels at Purdue University.