Ahead of the Curve: Engineering Simulation for Computers of the Future

引领潮流：未来计算机的工程仿真

基本信息

批准号：
EP/R04189X/1
负责人：
Steven Lind
金额：
$ 30.34万
依托单位：
University of Manchester
依托单位国家：
英国
项目类别：
Research Grant
财政年份：
2018
资助国家：
英国
起止时间：
2018 至无数据
项目状态：
已结题

来源：
https://gtr.ukri.org/projects?ref=EP%2FR04189X%2F1
关键词：
Ahead Curve Engineering Simulation Computers

项目摘要

Modelling and simulation is a key element of research, development and design in engineering generally; it assists in the creation and optimisation of tools and methods across industries from aerospace to manufacturing to biomedical. Despite the significant supercomputing power available today, we are still surrounded by problems of great societal and economic significance with a complexity far beyond the reach of modern computation. Examples include direct numerical simulation for turbulence on real-life scales (e.g. an entire aircraft), patient-specific digital drug trials, accurate weather prediction over months, rather than days. Emerging technologies and exascale supercomputers will greatly improve our simulation capability, and there are several EPSRC and EU funded projects addressing the challenges to simulation on these systems. In the not-to-distant future however, quantum computers promise to revolutionise the information age and, accordingly, the potential for simulating complex engineering problems. While exascale systems may calculate at a quintillion flops, quantum computers may be theoretically unlimited in clock speed, hindered instead by matters such as information transmission time. Practical quantum computers in mainstream use remain many years away, and while there is a concerted effort in hardware development (in research and industry), comparatively little attention is being paid to numerical algorithm development for engineering simulation on these new devices. This project addresses this oversight by exploring the amenability of popular contemporary numerical algorithms to quantum computing. Algorithm comparisons will be carried out on emulated quantum systems and, if possible, algorithm enhancements suggested for a better fit to future non-deterministic architectures. A key deliverable will be the first framework for practical numerical algorithm development for quantum machines for use by engineers in modelling and simulation in the coming decades.

建模和模拟是工程研究，开发和设计的关键要素。它有助于创建和优化从航空航天到制造业到生物医学的行业之间的工具和方法。尽管当今具有重大的超级计算能力，但我们仍然被社会和经济意义的问题所包围，复杂性远远超出了现代计算的影响力。示例包括直接数值模拟现实生活尺度（例如整个飞机），患者特定的数字药物试验，几个月内的准确天气预测，而不是几天。新兴技术和Exascale超级计算机将大大提高我们的模拟能力，并且有几个EPSRC和EU资助的项目解决了这些系统模拟的挑战。然而，在不遥远的未来，量子计算机承诺将彻底改变信息时代的年龄，并因此有可能模拟复杂的工程问题。尽管Exascale系统可以在千万亿次拖失板上计算，但理论上可能在时钟速度上是无限的，而是由于信息传输时间等问题而阻碍。主流使用中的实用量子计算机仍有多年的距离，尽管在硬件开发（研究和行业中）有一致的努力，但对这些新设备上的工程模拟的数值算法开发的关注很少。该项目通过探索流行的当代数值算法对量子计算的不适当性来解决这一监督。算法比较将在模拟的量子系统上进行，并在可能的情况下进行算法增强功能，建议更好地适合将来的非确定性体系结构。一个关键的可交付方式将是用于量子机的实用数值算法开发的第一个框架，用于工程师在未来几十年中用于建模和仿真中。