SI2-SSE: Development of a GPU Accelerated Gibbs Ensemble Monte Carlo Simulation Engine

该奖项支持开发使用低成本图形处理单元(GPU)进行加速的通用Gibbs Ensymble Monte Carlo(GEMC)模拟引擎。这项工作的主要目标是开发和实现：1)GPU加速配置偏差方法，2)在GPU上计算Ewald和的高效算法，3)针对不同GPU的代码自动调优。这项工作建立在PI的基础上？现有的基于颗粒的GPU-GEMC发动机，将引入能够模拟生物过程和多孔材料吸附的功能。编写代码将保持与NAMD和VMD软件包使用的文件格式的兼容性，从而简化模拟设置和数据分析。生成的模拟引擎将在GNU通用公共许可证v3(GPLv3)下发布，并通过互联网向用户提供。在该奖项的支持下开发的软件工具将实现二氧化碳封存材料的高通量计算筛选，用于稳定药物分散体的改进材料，并提供对膜融合等基本生物过程的分子水平的洞察。与传统的串行、受CPU限制的代码相比，使用图形处理器进行大部分计算工作预计将使计算时间减少一到两个数量级，这将允许模拟比现有工具更大规模和更复杂的系统。拟议的GPU加速蒙特卡罗模拟引擎的开发将增强生物化学、化学工程、材料科学和物理界的网络基础设施。GPU-GEMC模拟引擎将通过在国内和国际会议上的会议演讲、通过一个专门的网站以及通过发表同行评议的文献来推广。该奖项将加强研究生和本科生的教育。这项工作的研究主题将被整合到现有的GPU计算和分子模拟课程中。研究生和本科生将有机会成为由工程师和计算机科学家组成的多学科团队的一部分。学生将从传统上在STEM领域代表性较低的群体中招收。

This award supports the development of a general purpose Gibbs Ensemble Monte Carlo (GEMC) simulation engine that uses low-cost graphics processing units (GPU) for acceleration. The primary objectives of this work are to develop and implement: 1) GPU accelerated configurational-bias methods, 2) efficient algorithms for the computation of Ewald sums on the GPU and 3) automated tuning of the code for different GPUs. This work builds on the PIs? existing particle-based GPU-GEMC engine and will introduce functionality that enables the simulation of biological processes and adsorption in porous materials. The code will be written to maintain compatibility with the file formats used by the software packages NAMD and VMD, simplifying simulation setup and data analysis. The resulting simulation engine will be released under the GNU General Public License v3 (GPLv3) and made available to users via the Internet. The software tools developed with support from this award will enable high throughput computational screening of materials for CO2 sequestration, improved materials for the stabilization of drug dispersions, and provide molecular level insight to fundamental biological processes such as membrane fusion. The use of graphics processors for the bulk of the computational effort is expected to provide one to two orders of magnitude reduction in computational time compared to traditional serial, CPU bound code, which will allow for the simulation of systems of greater size and complexity than with existing tools. The development of the proposed GPU-accelerated Monte Carlo simulation engine will enhance the cyber-infrastructure of the biology chemistry, chemical engineering, materials science and physics communities. The GPU-GEMC simulation engine will be promoted through conference presentations at national and international meetings, via a dedicated website, and through publication in peer-reviewed literature.This award will enhance education at the graduate and undergraduate levels. Research topics from this work will be integrated into existing courses on GPU computing and molecular simulation. Graduate and undergraduate students will have the opportunity to work as part of a multidisciplinary team composed of engineers and computer scientists. Students will be recruited from groups traditionally underrepresented in STEM fields.