Benchmarking Quantum Advantage

The aim of this project is to investigate the practical advantage of quantum computing over contemporary classical computing. It will be conducted via simulations and modelling on high-performance clusters, based on the notion of incorporating quantum hardware in wider composite systems. In the end, we hope to design a framework for quantifying requirements needed to make such systems beneficial.

One of the biggest hurdles of modern quantum computing is feasibility. Unlike their classical counterparts, quantum bits (qubits) are extremely fragile and easily disturbed by noise, which severely hinders their performance and scalability. This means that we are currently in the era of noisy intermediate-scale quantum (NISQ) devices that support only a small number of qubits. This makes them interesting to study, as their advantage is also limited.

A standard technique to explore quantum systems is via simulations. It is usually assumed that simulated environments are noiseless, which is unfair compared to real hardware. However, it turns out that we can drastically reduce computation time and memory requirements by incorporating noise. This is done by leveraging tensor networks, which encode the dependencies between qubits (i.e. the entanglement).

In this project, we will perform large-scale noisy emulations of quantum systems to achieve significant improvements in running complex setups. The results will help us quantify the boundary beyond which quantum hardware achieves an advantage over available classical resources for a given problem. By inspecting the effects of advances in noise tolerance, and increasing the number of qubits, we can predict how the boundary is going to evolve, and how physical systems are likely to scale in the future. Ideally, we strive to design and develop a rigorous emulation framework for this purpose.

We will use C++ and Python programming languages to build the framework. It will be run on large computing clusters via parallel programming standards, such as MPI and OpenMP. The clusters to be used include ARCHER2 and Cirrus at EPCC. In addition, instead of writing quantum tensor networks from scratch, it may be possible to use existing libraries, and modify them to add support for noisy networks.

As the project progresses, we should be able to use the developed framework to investigate how quantum hardware integration can benefit high-performance computing (HPC). This is likely one of the first application of quantum computing, which can offer substantial speed-ups for certain applications, e.g. exponentially faster prime factorisation. Much work in this area is purely theoretical and assumes that a large number of error-corrected qubits is available, but we can expand it to near-term NISQ devices. Ultimately, we aim to construct and verify various models that would facilitate quantum integration into multi-component, heterogeneous architectures.

To sum up, the project has the following goals:
1. Simulate noisy quantum systems, eventually developing a framework for that purpose.
2. Evaluate the boundary between classical and quantum computing in the near-term.
3. Develop models for quantum hardware integration in high-performance systems.
Completing these will provide insight into the near future of quantum computing, which may be a crucial step towards its first practical implementations.