Development of a Parallel Reaction Path Package (ROW)

9212619 Baldridge这是向圣地亚哥超级计算机中心的Kim Baldridge博士颁发的女性研究机会奖(ROW)。计算化学是一个不断需要更快、更有效的工具来模拟化学反应的领域。随着越来越多的研究人员将实验与理论结合起来，以便更全面地阐明有关复杂反应的信息，这一点变得越来越重要。用半经验方法和从头算方法研究反应机理和预测反应速率常数受到计算反应势能面(PESS)所需计算量的限制。然而，在计算势面的能量梯度和高阶导数方面取得的进展是减少这一努力的一个重要因素，特别是通过允许有效地计算沿最小能量反应路径(MEP)的势。这就提出了一个问题，即为了获得足够收敛的反应路径表示以计算准确的热速率常数，必须以多高的精度来计算反应的MEP。答案在很大程度上取决于分子体系的大小和化学反应的复杂性。巴尔德里奇博士的建议在算法层面上解决了这些问题，并通过并行化在硬件层面上解决了这些问题。她建议在已经开展的工作的基础上再接再厉，为并行计算环境完成急需的计算化学程序包(MOPAC)的再开发。MOPAC是一种广泛使用的半经验量子化学程序，用于研究化学反应。

9212619 Baldridge This is a Research Opportunities for Women (ROW) award to Dr. Kim Baldridge at the San Diego Supercomputer Center (SDSC). Computational chemistry is an area constantly in demand of faster, more efficient tools to model chemical reactions. This is of growing importance as more and more researchers meld experiment with theory in order to elucidate more fully information concerning complex reactions. The study of reaction mechanisms and the prediction of rate constants by semi-empirical and ab initio methods is limited by the computational effort needed to calculate reactive potential energy surfaces(PESs). The advances made in the calculation of energy gradients and higher order derivatives of potential surfaces however, have been an important factor in reducing this effort, in particular by allowing efficient calculation of the potential along a minimum energy reaction path(MEP). This raises the question of the accuracy with which the MEP for a reaction must be calculated in order to obtain a reaction path representation well enough converged to calculate accurate thermal rate constants. The answer highly depends on the size of the molecular system and the complexity of the chemical reaction. Dr. Baldridge's proposal addresses these problems at the algorithmic level and, via parallelization, at the hardware level. She proposes to build on work already undertaken in order to complete the redevelopment of a much needed computational chemistry package (MOPAC) for the parallel computing environment. MOPAC is a widely used semi- empirical quantum chemistry code for the study of chemical reactions.