Many scientific applications consist of heavy computational and analysis workload on data, and often require producing intermediate data for ongoing calculations. For instance, chemical dynamics simulations are known as heavy workload applications in terms of calculation in many aspects. There is a strong desire of seeking a solution to minimize expensive calculations by replacing them with light-weight ones. VENUS is one of these chemical dynamic simulation software packages known as classical chemical dynamics simulation, with scalar executing code and heavy calculation process. In this research, we introduce an innovative approximation method by storing, managing, and leveraging intermediate data (results) in order to speed up expensive calculations. The key idea is a newly introduced data interpolation method that leverages data points from previous calculations. The newly proposed method is a general approach that can be applied to a variety of scientific applications and disciplines. In this research, we focus on chemical dynamics simulations and the VENUS code and have developed a prototype of the data interpolation method for reduced computations. The proposed computation reduction method through increased data re-use can increase the efficiency and productivity of scientific simulations, thus can have an impact on scientific discovery powered by high performance computing simulations.

许多科学应用程序都需要对数据进行繁重的计算和分析工作，并且经常需要为进行中的计算生成中间数据。例如，就许多方面而言，化学动力学模拟在计算方面被称为繁重的工作量应用程序。强烈希望寻求一种解决方案，以轻量级计算代替昂贵的计算。VENUS是这些化学动力学仿真软件包之一，被称为经典化学动力学仿真，具有标量执行代码和繁重的计算过程。在这项研究中，我们通过存储，管理和利用中间数据（结果）引入了一种创新的近似方法，以加快昂贵的计算速度。关键思想是新引入的数据插值方法，该方法利用了先前计算中的数据点。新提出的方法是一种通用方法，可以应用于各种科学应用和学科。在这项研究中，我们专注于化学动力学模拟和VENUS代码，并开发了用于简化计算的数据插值方法的原型。通过增加数据的重用性，提出的减少计算方法可以提高科学模拟的效率和生产率，从而可以对高性能计算模拟提供动力的科学发现产生影响。我们专注于化学动力学模拟和VENUS代码，并开发了用于简化计算的数据插值方法的原型。通过增加数据的重用性，提出的减少计算方法可以提高科学模拟的效率和生产率，从而对以高性能计算模拟为动力的科学发现产生影响。我们专注于化学动力学模拟和VENUS代码，并开发了用于简化计算的数据插值方法的原型。通过增加数据的重用性，提出的减少计算方法可以提高科学模拟的效率和生产率，从而可以对高性能计算模拟提供动力的科学发现产生影响。