Quantum Enhanced and Verified Exascale Computing - QEVEC

鉴于计算硬件的先进能力，我们需要开发能够充分挖掘其潜力的能力。这对各种类型的专业计算硬件（如gpu（图形处理单元）或量子计算机）来说是一个挑战。最有效地结合不同类型的计算硬件来解决复杂问题对于下一代高性能计算至关重要，这将达到百亿亿次的处理速度（每秒超过10^18次操作）。QEVEC——量子增强和验证百亿亿次计算——通过专注于如何将量子计算机作为协处理器添加到传统高性能计算机（HPC）中，为这一发展做出了贡献。早期的量子计算机将比目前的HPC小得多——就它们一次可以处理的经典数据量而言。但由于它们的量子叠加和相干特性，对这些数据的处理能力可以快得多。因此，最有希望使用它们的方法是加速那些对HPC来说很慢的计算部分。这需要对量子和经典算法进行详细的研究，QEVEC将为流体模拟和材料模拟这两个特定的应用领域进行研究。流体模拟需要巨大的计算能力来求解非线性微分方程，通常是在复杂的几何形状和几种不同性质的流体的混合物中。有一系列的算法可以做到这一点，它们也可以作为非线性微分方程求解器有更广泛的应用。重要的应用范围包括天气预报，制造过程中的流体流动，聚变反应堆和恒星的等离子体模拟。材料模拟直接遇到了用经典计算机模拟量子系统的困难。量子系统具有比经典系统大得多的状态空间，这需要在经典计算机上使用大量的内存。因此，量子计算机是试图提高性能的自然选择。已经有很多小组为此研究变分量子算法；QEVEC将专注于更先进的方法来模拟固溶体（由两种化合物组成的金属合金和其他材料，它们单独或组合时采用相同的晶体结构），以及支持所有生物分子、化学和材料模拟的第一性原理的基本电子结构计算。有必要在某种程度上检查所有的计算机计算，以了解我们可以在多大程度上相信它们是正确的。量子计算机需要一些额外的技术来完成这种检查，因此QEVEC还将开发用于流体和材料模拟应用的量子验证方法。为了确保QEVEC团队的研究成果能够广泛地供科学计算社区使用，QEVEC将与其他ExCALIBUR项目团队广泛合作，分享他们的知识，并向其他项目学习。

Given the advancing capabilities of computing hardware, we need to develop capabilities to extract its potential fully. This is made challenging by various types of specialist computing hardware, such as GPUs (graphical processing units), or quantum computers. Combining different types of computing hardware most productively to solve complex problems is crucial for the next generation of high performance computing that will attain exascale processing speeds (more than 10^18 operations per second).QEVEC - quantum enhanced and verified exascale computing - contributes to this development by focusing on how to add quantum computers as co-processors to conventional high performance computers (HPC).Early quantum computers will be much smaller -- in terms of the amount of classical data they can process in one go -- than current HPC. But the processing power on that data can be much faster due to their quantum properties of superposition and coherence. The most promising way to use them is thus to accelerate those parts of the computations that are slow for HPC.This requires detailed study of the algorithms, both quantum and classical, which QEVEC will do for two specific applications areas, fluids simulations and materials simulations.Fluids simulations require immense computing power to solve the nonlinear differential equations, often in complex geometries and with mixtures of several fluids with different properties. There are a range of algorithms that can do this, and they can also have wider applications as nonlinear differential equation solvers. Important applications range through weather forecasting, fluid flows in manufacturing processes, plasma simulations for fusion reactors and stars.Materials simulations come directly up against the difficulty of simulating quantum systems with classical computers. Quantum systems have a much larger state space than classical systems, which requires a huge amount of memory on classical computers. A quantum computer is thus a natural choice to try to improve performance. There are already many groups studying variational quantum algorithms for this; QEVEC will focus on more advanced methods for simulating solid solutions (metal alloys and other materials composed of two compounds that individually and when combined adopt the same crystal structure), and the basic electronic structure calculations that underpin all first principles biomolecular, chemical and materials simulations.It is necessary to check all computer calculations at some level, to know how well we can trust them to be correct. Quantum computers need some extra techniques to accomplish this checking, so QEVEC will also develop quantum verification methods for the fluids and materials simulations applications.To ensure the results from the research carried out by the QEVEC team are widely available for the scientific computing community to benefit from, QEVEC will engage widely with other ExCALIBUR project teams to share their knowledge, and learn from the other projects.