Enabling next generation quantum chemistry

这一建议解决了在电子结构的第一性原理计算方面取得进展的需要，以便在几个化学重大挑战上取得进展。目前，有些问题是完全无法解决的；这些包括:。使用最佳电子波函数方法生成生物分子的第一性原理力场。整个1000原子金属纳米颗粒表面的催化作用超越时间依赖密度泛函理论的光活性生物学。溶剂化体系和溶液中化学反应的预测光谱。传统上，量子化学代码非常复杂，需要许多人的时间来开发和后续维护。多年来，这些软件系统一直以一种特殊的方式开发，受到快速表达新科学的需要的驱动，并且通常没有为单个最佳实现抽象关键的通用组件。不幸的是，这给进步带来了障碍。在数据和任务管理不抽象的情况下，新的高性能计算系统难以开发。新理论和新算法的实现受到当前代码复杂性的阻碍。此外，不同编码的多种输出模型阻碍了后处理和预处理；一个统一的可扩展模型会更好。本提案是一份初步旅行资金申请，旨在维持和加强英国和美国研究人员之间的联系，该项目是在6月13日至15日在牛津举行的EPSRC-NSF联合研讨会上建立的。

This proposal addresses the need for advances to be made in the first-principles computation of electronic structure in order for progress to be possible on several of the chemical grand challenges. At present, some problems are completely inaccessible; these include:. First-principles force field generation for biomolecules using the best electronic wavefunction methodology.. Catalysis on the surface of entire 1000-atom metallic nanoparticles photoactive biology beyond time-dependent density functional theory.. Predictive spectroscopy of solvated systems and chemical reactions in solution.Traditionally, quantum chemistry codes are extremely complex, and require many man-years for development and subsequent maintenance. Over the years, these software systems have been developed in an ad-hoc way, driven by the need to express new science quickly, and often without abstraction of key generic components for single optimal implementation. Unfortunately, this presents barriers for progress. New HPC systems are difficult to exploit when data and task management are not abstracted. Implementation of new theories and new algorithms is held back by present code complexity. Furthermore, the multiple output models of different codes hinder post- and pre-processing; a single agreed extensible model would be better.This proposal is an application for preliminary travel funding in order to maintain and strengthen the links between the UK and US researchers for this project that were established in the joint EPSRC-NSF workshop that was held in June 13-15 in Oxford.