MRI: Acquisition of a high-performance computer cluster for the computational study of complex chemical systems: from small molecules to biological nanomachines

理论化学中最好的工作结合了新数学方法的创新创造，这些方法的计算实现的新算法的发展，以及对化学中最广泛意义上的关键系统的大规模，计算机密集型模拟。马里兰大学的米勒德·亚历山大、约翰·威克斯、丹尼尔·科索夫、克里斯托弗·雅兹恩斯基和德夫·蒂鲁马莱在这三个领域都表现出色。他们的工作涵盖了最广泛意义上的化学和化学生物学，包括小分子动力学中的非绝热效应（Alexander），复杂系统的热力学性质（Weeks），通过纳米结构和分子的非平衡电子传递（Kosov），复杂系统的热力学性质（Jarzynski）和核糖体组装机制（Thirumalai）。主要研究人员已经在这些领域建立了变革性的前沿研究项目。此外，几位主要研究人员在他们各自的领域为开发最先进的软件包做出了贡献，这些软件包现在正在马里兰大学以外的许多研究小组中使用。理论化学和化学生物学的一个主要挑战是将精确的建模扩展到更大的复杂性和维度的系统。主要研究人员目前可用的计算资源要么过时，要么过饱和，因此不足以进行研究和培训活动。目前的MRI提案将允许购买一个基于双核、多cpu服务器节点的新集群。从拟议的研究中获得的这一见解将在化学和化学生物学的许多领域有潜在的应用，从燃烧到大气化学，到复杂流体和界面，到纳米材料中的电流流动，到复杂系统中热力学平衡的理解以及生物纳米机器的功能。所有这些领域对我们整个社会都至关重要，特别是对发展更清洁的燃烧、更有效的润滑、更高通量的流体输送、新的纳米材料和更好地理解基本的生物转化。此外，所提出的研究将导致理论方法和计算代码的持续发展，以实现上述领域的准确，快速模拟，利用微观和宏观并行化。这些代码将免费分发给整个科学界。此外，未来的研究科学家，本科生，研究生和博士后研究员将在使用计算化学和生物物理学的最新工具以及高效使用高性能，大规模并行计算机架构的开发策略方面进行培训。这些技能对于确保美国的技术领先地位至关重要。

The best work in theoretical chemistry combines innovative creation of new mathematical methods, the development of new algorithms for the computational implementation of these methods, and large-scale, computer-intensive simulations on systems of key importance in chemistry in its broadest sense. Excellence in all three of these area characterizes the work of Millard Alexander, John Weeks, Daniel Kosov, Christopher Jarzynski, and Dev Thirumalai of the University of Maryland. Their work covers chemistry and chemical biology in the broadest sense, ranging from including non-adiabatic effects in small molecule dynamics (Alexander), thermodynamic properties of complex systems (Weeks), non-equilibrium electron transport through nanostructuctures and molecules (Kosov), thermodynamic properties of complex systems (Jarzynski), and the mechanism of ribosome assembly (Thirumalai). The principal investigators have already established transformational, forefront research programs in these areas. In addition, several of the principal investigators have contributed to the development of state-of-the-art software packages in their individual areas which are now being used in many research groups external to the University of Maryland. A major challenge in theoretical chemistry and chemical biology is the extension of accurate modeling to systems of larger complexity and dimensionality. The computational resources currently available to the principal investigators are either outdated or oversaturated, and thus have become insufficient for their research and training activities. The present MRI proposal will allow purchase of a new cluster, based on dual-core, multi-cpu server nodes. This insight gained from the proposed research will have potential application in many areas of chemistry and chemical biology, ranging from combustion, to atmospheric chemistry, to complex fluids and interfaces, to current flow in nanomaterials, to the understanding of thermodynamic equilibrium in complex systems and to the functioning of biological nanomachines. All of these areas are of crucial importance to our society as a whole, in particular to the development of cleaner combustion, more efficient lubrication, higher throughput fluid transport, new nano-materials and for a better understanding of fundamental biological transformations. Also, the proposed research will lead to the continued development of theoretical methods and computational codes for accurate, fast simulations in the areas described above, making use of both micro- and macro-parallelization. These codes will be freely distributed to the scientific community at large. In addition, future research scientists undergraduates, graduate students and postdoctoral fellows will be trained in the use of the latest tools in computational chemistry and biophysics and in the development of strategies for efficient use of high-performance, massively parallel computer architectures. These skills are crucial to ensure the technological leadership of the United States.