SCREMS: High order numerical algorithms and their applications

该奖项为保障计算机设备的初始设置和维护提供资金，以支持几个NSF支持的项目的计算需求，包括对流主导问题的算法设计和高阶数值方法的应用，弹性和数值相对论的高阶数值方法。正在研究的算法包括有限差分和有限体积WENO格式、不连续Galerkin和其他有限元方法以及谱方法，特别是在自适应、多尺度和不确定环境中。一些建议工作的特别重点是开发和高阶方法的算法适应图形处理单元（GPU）加速架构。在这个项目中开发的算法将有助于提高我们理解许多重要问题的基本物理的能力，并促进先进材料、飞机、半导体器件和能源产品的工程设计和生产。

This award provides funds to secure computer equipment with initial setup and maintenance to support the computational requirements of several NSF supported projects, including algorithm design and application of high order numerical methods for convection dominated problems and high order numerical methods for elasticity, and numerical relativity. The algorithms being investigated include the finite difference and finite volume WENO schemes, discontinuous Galerkin and other finite element methods, and spectral methods, especially in adaptive, multiscale and uncertain environments. A particular emphasis of some of the proposed work is development and algorithmic adaptation of high-order methods to graphics processing unit (GPU) accelerated architectures. The algorithms developed in this project will help to improve our capability to understand the fundamental physics of many important problems, and to facilitate the engineering design and production of advanced materials, aircrafts, semiconductor devices, and energy products.