Castep: Advanced spectroscopies using high-performance computing

随着最近理论和计算的进步，我们已经能够从第一性原理计算凝聚态系统的性质。人们甚至可以希望做到这一点，这要归功于量子力学在描述化学键时的准确性。狄拉克的一句“发现量子力学之后剩下的就是化学了”的伪讽刺总结了这句话：如果一个人能解出某种东西的薛定谔方程——一个原子、一个分子、固体或液体中的原子组合——他就能预测每一种物理性质。狄拉克的陈述并没有显示出做剩下的事情有多么困难，它需要我们付出巨大的努力和聪明才智，才能以合理的精度计算出材料的某些性质。模拟对我们思考凝聚态问题的影响是巨大的。然而，CASTEP代码试图仅使用量子力学（特别是密度泛函理论）来计算材料的许多特性。CASTEP是一个功能丰富的第一性原理电子结构代码，因此它的功能是众多的。旨在从第一性原理计算系统的任何物理性质，基本量是由许多其他量派生的总能量。在这里，我们希望进一步发展CASTEP的应用范围，这对更广泛的科学家来说是有价值的，这将使他们能够以快速，可靠和准确的方式进行研究。它是专门为高性能计算机使用而设计的，从底层开始编写时就考虑到了多核架构。Castep是英国超级计算设备HECToR中最常用的代码之一。我们的目标是开发一系列令人兴奋的新光谱工具，这些工具将在理论/计算凝聚态物质和实验技术之间形成紧密的联系。解释实验光谱数据并不简单，但如果建立在坚实、准确的理论基础上的软件可以用来生成这样的数据，那么就可以直接与实验进行比较，并允许对结果进行详细的解释。这就是当前提案的目的。在整个过程中，我们将保持最高质量的代码设计/实现和测试技术，就像我们过去一贯做的那样。在这项工作的最后，我们将在CASTEP中拥有新的功能，这将产生新的科学。一个进一步的提议（这个号召的第2阶段）将被用来使代码达到一个可以被广泛使用的质量。

With recent theoretical and computational advances we have been able to calculate the properties of condensed matter systems from first principles. That one can even hope to do this is down to the accuracy of quantum mechanics in describing the chemical bond. Dirac's apocryphal quip that after the discovery of quantum mechanics the rest is chemistry sums it up: if one can solve the Schrodinger equation for something - an atom, a molecule, assemblies of atoms in solids or liquids - one can predict every physical property. Dirac's statement doesn't quite show how difficult doing the rest is, and it has taken great effort and ingenuity to take us to the point of calculating some of the properties of materials with reasonable accuracy. The impact of simulations on our thinking about condensed matter problems is immense. However, the CASTEP code attempt to calculate many properties of materials using only quantum mechanics (in particular, density functional theory).CASTEP is a richly featured first principles electronic structure code and as such its capabilities are numerous. Aiming to calculate any physical property of the system from first principles, the basic quantity is the total energy from which many other quantities are derived. Here we wish to develop CASTEP's range of applications further, making is valuable to a much wider range of scientists which will enable them to perform their research in a fast, reliable and accurate manner.It has been designed specifically for use on high-performance computers, being written from the ground up with a many-core architecture in mind. Castep is one of the most used codes on the UK supercomputing facility, HECToR. We aim to develop an exciting range of new spectroscopic tools that will form a close link between theoretical/computational condensed matter and experimental techniques. Interpretation of experimental spectroscopic data is not straightforward, but if software that is build on a firm, accurate theoretical foundation can be used to generate such data, then direct comparison to experiment can be performed and allow detailed interpretation of the results to be done. This is the aim of the current proposal. Throughout we will maintain the highest quality code design/implementation and testing techniques, as we have consistently done in the past. At the end of this work we will have new functionality in CASTEP that will produce new science. A further proposal (stage 2 of this call) will be used to make the code of a quality such that it can be widely used.