Simulators and HPC’s Role in the NISQ Era

Now that we know how to make quantum and classical computing resources available and have reviewed how to pose our problems in both domains, we should evaluate the available mechanisms and strategies for exploiting those resources efficiently. By that, we mean cost and time efficiency, given that those axes will also need to be considered when it comes to including these techniques in our company’s daily processes.

Nowadays, the classical resources in most companies comprise a mixture of on-premises and cloud-enabled resources. This is the common case for most experimental projects aiming to improve operational processes using analytics. Ephemeral computing resources may have different needs, depending on the project or the nature of the technique we envision using. That is why the cloud-native pay-per-use model has become a good option for most companies.

Depending on the tasks, graphical processing units (GPUs) for machine...

Local simulation of noise models

First, we must distinguish between the three naming conventions we will use during this chapter and the following ones. Previously, we talked about how quantum algorithms can be run on a classical device before being sent to a real quantum device, but there are some different ways in which this classical execution can be done. Problems such as the ones we have posed have been around for a while, and classical computing has evolved in many different ways to bring solutions to the technology at hand during this time. As we will see, this may also bring some challenges related to the specifics of the different classical setups we will cover. Mimicking quantum mechanical evolution is a non-trivial task; that was how quantum computing was proposed as a potential solution.

Simulators are the classical means of processing information in the way an ideal quantum computer would do so. Recall that quantum information theory is not a new task brought about...

Summary

In this chapter, we saw that there are many ways to simulate a quantum computer before running it on an actual device. We saw that there are also some implications regarding the limited availability, errors, and specific characteristics of the real hardware to be considered and that classical computers are not yet done when it comes to quantum computing.

Establishing a strategy to validate our circuits, evaluate their potential, and decide where those algorithms will run requires understanding the set of options provided by almost all quantum companies.

Tensor networks provide a powerful mathematical framework to simulate complex systems efficiently. GPUs have also placed their bet. Even combining both has proven to be a valid approach for simulating large devices.

Distributed computation is anticipated to be the next hurdle to overcome, necessitating a certain level of technical expertise to harness its potential efficiently. Similar to the trajectory followed by...

Further reading

For those interested in diving deeper into some of the techniques mentioned in this chapter, here are some recommendations that should help you understand the basics.

One of the most interesting and challenging frameworks we have discussed is tensor networks. Many resources can be found in the literature. Still, two that we can recommend are the work by Biamonte and Bergholm from 2017, which provides a solid foundation to understand its potential better. For those more hands-on engineers, the Quimb (Gray, 2018) and Jet (Vincent et al., 2022) Python packages provide a fun way to learn and experiment.

Similarly, distributed computation has a path, and works by Zaharia et al. (2010) on Apache Spark and Moritz et al. (2018) on Ray are leading the path toward easy-to-implement distributed solutions.

Something particularly interesting is the contribution of the Baidu team to the existing PaddlePaddle framework (Ma et al., 2020). Not only have they provided an industrial...

References

Bennett, C. H., & Shor, P. W. (1998). Quantum information theory. IEEE transactions on information theory, 44(6), 2,724-2,742.

Biamonte, J., & Bergholm, V. (2017). Tensor networks in a nutshell. arXiv preprint arXiv:1708.00006.

Brassard, G., Hoyer, P., Mosca, M., & Tapp, A. (2002). Quantum amplitude amplification and estimation. Contemporary Mathematics, 305, 53-74.

Bowen, G. (2001). Classical information capacity of superdense coding. Physical Review A, 63(2), 022302.

Guerreschi, G. G., Hogaboam, J., Baruffa, F., & Sawaya, N. P. (2020). Intel Quantum Simulator: A cloud-ready high-performance simulator of quantum circuits. Quantum Science and Technology, 5(3), 034007.

Gray, J. (2018). quimb: A Python package for quantum information and many-body calculations. Journal of Open Source Software, 3(29), 819.

Kwak, Y., Yun, W. J., Kim, J. P., Cho, H., Park, J., Choi, M., ... & Kim, J. (2022). Quantum distributed deep learning architectures...