MRI: Acquisition of a High Performance Computing System for Science and Engineering Research and Education at the University of Houston

该主要研究仪器（MRI）项目提供了一台高性能计算（HPC）仪器，以支持休斯顿大学（UH）系统以及休斯顿地区其他少数民族服务机构的研究。它使跨学科的研究工作：多尺度生物分子模拟；原子分子动力学模拟；替代能源的计算化学；气相、凝聚相和纳米结构材料的模拟；以及基于多核和图形处理单元（GPU）的系统的大型程序重构。该设施还开发了早期宇宙宇宙学的模拟；随机网络仿真；粒子物理中的量子色动力学晶格模拟；空间辐射分析与模拟；波函数的地震表征；进化的模拟;基因组数据的计算分析；云计算和图像处理。该仪器显著增加了参与组织的科学、技术、工程和数学（STEM）学科研究人员可用的资源。校园内对异构计算平台的访问促进了计算科学家和编程专家之间的协作，以推进图形处理单元（gpu）在科学计算中的使用。开发培训材料和策略以支持代码开发并充分利用性能优势。新一代开源的、基于指令的、支持gpu的应用程序代码被创建出来，展示了可伸缩性和可移植性。该项目极大地丰富了大学的研究基础设施，并增加了休斯顿大学、西班牙裔服务机构和合作机构对新型高性能计算技术和系统的使用。UH与两所历史上的黑人学院（HBCUs）合作，扩大来自代表性不足群体的学生在STEM学科中的参与。这种影响对本科生的研究尤其明显。学生和研究人员接触到充满活力和多学科的研究环境，并接受有价值的高级HPC培训-这些技能很容易在科学领域之间转移。在该奖项下开发的算法、计算机代码和数据集的各个方面都是公开的。

This Major Research Instrument (MRI) project provides for the acquisition of a high-performance computing (HPC) instrument to support research from the University of Houston (UH) system as well as other minority-serving institutions in the Houston region. It enables cross-disciplinary research efforts in: multiscale biomolecular simulations; atomistic molecular dynamics simulations; computational chemistry for alternative energy; simulations of gas-phase, condensed-phase and nanostructured materials; and large-program restructuring for multicore and graphics processing unit (GPU) based systems. This facility also develops simulations of early universe cosmology; simulations of stochastic networks; quantum chromodynamics lattice simulations for particle physics; space radiation analysis and simulations; seismic characterization of wave functions; evolutionary simulations; computational analysis of genomic data; and cloud computing and image processing. The instrument significantly increases resources available to researchers in science, technology, engineering, and mathematics (STEM) disciplines at the participating organizations. On-campus access to a heterogeneous computing platform fosters collaboration between computational scientists and programming experts to advance the use of graphics processing units (GPUs) for scientific computing. Training materials and strategies are developed to support code development and fully exploit the performance benefits. A new generation of open-source, directive-based GPU-enabled application codes are created that demonstrate scalability and portability. This project greatly enriches the university research infrastructure and increases the use of new HPC techniques and systems at the University of Houston, a Hispanic Serving Institution, and collaborating institutions. UH partners with two Historically Black Colleges (HBCUs) to broaden the participation of students from underrepresented groups in STEM disciplines. The impact is particularly pronounced for undergraduate research. Students and researchers are exposed to a vibrant and multi-disciplinary research environment and receive valuable, advanced training in HPC - skills that are readily transferrable between scientific domains. Aspects of algorithms, computer codes, and data sets developed under this award are made publicly available.