Quantum computers are domain-specific accelerators that can solve commercially important problems that are beyond the capability of conventional computers. Quantum computing is at an inflection point where small systems with a few dozen qubits have been demonstrated and the number of qubits is expected to increase to several hundred over the coming years. However, the qubit quality is likely to remain quite low making it difficult to execute programs reliably. Furthermore, scaling the designs to a large number of qubits would require specialized hardware that operates at cryogenic conditions. This article describes the compiler and hardware support for enabling reliable and scalable quantum computers.
Quantum computers are domain-specific accelerators that can solve commercially important problems that are beyond the capability of conventional computers. Quantum computing is at an inflection point where small systems with a few dozen qubits have been demonstrated and the number of qubits is expected to increase to several hundred over the coming years. However, the qubit quality is likely to remain quite low making it difficult to execute programs reliably. Furthermore, scaling the designs to a large number of qubits would require specialized hardware that operates at cryogenic conditions. This article describes the compiler and hardware support for enabling reliable and scalable quantum computers.
Quantum computing as emerging technology raises high expectations to accelerate complicated calculations, superseding capabilities of presently available supercomputers and to provide a solution to the ever-increasing challenge of processing big amount of data. Besides theoretical promises, quantum computers have successfully been realized in experimental settings during the past few years, and can now also be accessed commercially. Given this availability of quantum computers, their application potential can be scrutinized. Promised to lead to disruptive changes in addressing and solving problems from the domains of simulation, optimization and machine learning, quantum computers have been researched in context of many application scenarios in industrial and academic setting in the recent years. However, it turns out that currently available quantum computers are still to small in size and limited in quality to fulfill the high hopes set in them at the moment. This review reflects on the current state of the technology coming from the perspective of the application, and indicates the developments required on hardware, software and algorithmic side to eventually make quantum computing beneficial for applications.
Quantum computers offer the promise of accelerated solutions for types of problems that are computationally unaffordable or intractable for classical computers, such as the simulation of quantum mechanics itself, quantum chemistry, and mathematical challenges like factoring large numbers. However, the relatively high error rate per quantum computing operation in the systems of today and of those expected to be developed in the future remains a key obstacle. To address this challenge, quantum error correction (QEC), which enables the creation of error-corrected (logical) qubits from a collection of physical qubits, is required. Depending on the choice of QEC code, hundreds of physical qubits may be required to implement a single logical qubit. As a result, the race to the creation of quantum computing systems capable of outperforming classical computing systems is largely a race to create quantum computers at the scale needed to support QEC. The criticality of scaling is evident in the quantum computing roadmaps produced by major research labs in this field, such as the one from IBM shown in Fig. 1. Starting from systems with only a few physical qubits a few years ago, the roadmap projects quantum computers with around 100,000 physical qubits and thousands of logical qubits by 2033 or so.
It is expected that a useful quantum computing (QC) system would require millions of qubits within one system. The state-of-the-arts QC systems contain hundreds of qubits, which is still several orders of magnitude away from the ultimate goal. To scale the QC system, namely, to increase the number of qubits and QC circuit layers, and also to improve the performance such as fidelity, robustness, etc., we need to develop scalable techniques including high-density qubit devices, ultra-low power qubit interfacing circuits, 3D interconnection and packaging etc. This article will review the recent reported device, circuit and system-level techniques that contribute to the QC system scaling and may ultimately lead to a fully functional quantum computer.
For the realization of a quantum computer, a good qubit design as the basic building block is a nontrivial starting point. To fully leverage the potential of quantum computing, quantum processors will likely have to incorporate billions of high-quality qubits. Several approaches are being pursued and exhibit different strength and weaknesses. The device concepts and fabrication procedures for semiconductor based qubits are very similar to those used in the semiconductor industry. Using the spin-degree of freedom, their coherence properties and the quality of single-qubit operations have now reached the required level and are on par with competing approaches. Two-qubit operations will likely follow suit in the near future. With these basic requirements being met, it is now time to think if and how increasingly large multi-qubit systems and eventually functional quantum processor can be realized. This quest raises many pertinent questions, some of which equally apply to other qubit platforms. How reproducibly can the different types of semiconductor qubits be produced? How can their operating point be adjusted efficiently and in an automated way? How can they be connected to form meaningful multi-qubit circuits? How severe are multi-qubit control issues such as crosstalk and frequency crowding, and how can they be avoided? What control system architectures are most suitable for operating large numbers of qubits? What is the limit of cryogenic electronics regarding power consumption and integration density? This paper gives a (partial) overview of the state of the art in the field and discusses future challenges associated with the realization of large-scale quantum processors, many of which will greatly benefit from input from engineering science. The guiding vision is that of a highly integrated quantum processor with an ultra-low power cryogenic control system.
Despite numerous advances in the field and a seemingly ever-increasing amount of investment, we are still some years away from seeing a production quantum computer in action. However, it is possible to make some educated guesses about the operational difficulties and challenges that may be encountered in practice. We can be reasonably confident that the early machines will be hybrid, with the quantum devices used in an apparently similar way to current accelerators such as FPGAs or GPUs. Compilers, libraries and the other tools relied upon currently for development of software will have to evolve/be reinvented to support the new technology, and training courses will have to be rethought completely rather than ``just'' updated alongside them. The workloads we are likely to see making best use of these hybrid machines will initially be few, before rapidly increasing in diversity as we saw with the uptake of GPUs and other new technologies in the past. This will again be helped by the increase in the number of supporting libraries and development tools, and by the gradual re-development of existing software, to make use of the new quantum devices. Unfortunately, at present the problem of error correction is still largely unsolved, although there have been many advances. Quantum computation is very sensitive to noise, leading to frequent errors during execution. Quantum calculations, although asymptotically faster than their equivalents in ``traditional'' HPC, still take time, and while the profiling tools and programming approaches will have to change drastically, many of the skills honed in the current HPC industry will not suddenly become obsolete, but continue to be useful in the quantum era.
Although the concept of quantum computing has existed for decades, the technology needed to successfully implement a quantum computing system has not yet reached the level of sophistication, reliability, and scalability necessary for commercial viability until very recently. Significant progress on this front was made in the past few years, with IBM planning to create a 1000-qubit chip by the end of 2023, and Google already claiming to have achieved quantum supremacy. Other major industry players such as Intel and Microsoft have also invested significant amounts of resources into quantum computing research. Any viable computing system requires both hardware and software to work together harmoniously in order to perform useful computations. While the achievements of IBM and other companies represent a large step forward for quantum hardware, many gaps remain to be filled with respect to the corresponding software. Specifically, there is currently no clear path towards a complete process for translating quantum algorithms into physical operations that are directly executable on quantum hardware. Such a process is analogous to a compiler that translates programs written in a high-level language into executable machine instructions on a conventional digital computer, and it is necessary if quantum computers are to be harnessed to perform practically useful computations. Existing work has addressed individual components of this process, but so far no unified method for translating the whole of a quantum algorithm into executable operations has been described. I make substantial progress towards filling this gap by describing a set of high-level and low-level quantum circuit design techniques, which when taken together reduce the need of a circuit designer to be concerned with low-level details. On the high-level side, I describe an approach or strategy to designing quantum oracles for Grover's algorithm that allows it to be applied to several types of problems. This approach involves designing oracles in terms of high-level blocks such as counters, multiplexers, comparators, and arbitrary Boolean functions. The implementations of these blocks in terms of lower-level quantum gates are demonstrated in a way that makes it clear that scaled-up versions of them can be generated in a completely automated fashion. For a specific class of problems, which I call state-space path planning problems, I also introduce a paradigm for quantum oracle design that involves representing the problem in terms of individual states and moves. Problems of this sort have applications in robotics and games. Low-level techniques that I introduce