This research was partly supported by a grant under the Japan Science and Technology Agency (JST) Moonshot Research and Development Program Goal 6, “Realization of a fault-tolerant universal quantum computer that will revolutionize economy, industry, and security by 2050” (Program Director Masahiro Kitagawa), for the R&D Project “Large-scale Silicon Quantum Computer” (Project Manager Hiroyuki Mizuno; grant number JPMJMS2065).
A shift from software research to the world of quantum computers
Sato: In university, my research was in software engineering. I actually preferred the humanities, and enjoyed my classes in history, political economy and other social studies; but since I did well in science and technology subjects, I decided to take up work matched to my abilities. This was an easy decision, knowing that I would always be able to enjoy topics I liked, even without making a career of them. My specific research in the software engineering field in university was on formal verification techniques.
A reason for choosing to join Hitachi is that the company has many different domains, and a broad range of areas for applying software engineering. Previously, I have conducted research on software for assisting developers of systems in the railway field and in the finance field. In the railway field, for example, I studied ways of raising program development productivity by having programs generated automatically, simply by creating figures and tables. I believe this was at the forefront of what is today called the no-code or low-code movement. My involvement in the silicon quantum computer R&D project began in 2021.
Miyamoto: My major in university was in computer science, and in graduate school I studied in the field of computer and mathematical sciences. For our first seminar exercise, we were handed a copy of Computer Organization and Design by Patterson and Hennessy, a classic reference work on computers. It was great fun actually trying my hand at design with that as a handbook. Thereafter, I continued my research on encrypted circuit design and obtained my doctoral degree.
Wanting to take on challenges in new fields, I went on to choose Hitachi as my place of work, with its broad domains. Upon entry, I was assigned to a computing-related division headed by Hiroyuki Mizuno, who is today a Distinguished Researcher leading the quantum computer research. There I was involved in research on computing for large-scale graph processing. Since my specific focus was on application-layer software aimed at large-scale graph processing, I gradually shifted from hardware to software research. Right before I was transferred to the quantum computer project, I was researching AI technology including machine learning.
The technology theme this time is “a shuttling qubit method”
Miyamoto: In fall of 2020, around a half year before Sato-san, I became part of a research unit carrying on silicon quantum computer studies. Around that time, Hitachi had been selected for the “Large-scale Silicon Quantum Computer” project that is one of the R&D projects in the government’s Japan Science and Technology Agency (JST) Moonshot Research and Development Program, Goal 6, “Realization of a fault-tolerant universal quantum computer that will revolutionize economy, industry, and security by 2050” (Program Director Masahiro Kitagawa, professor in the Graduate School of Engineering Science, Osaka University). I was asked to join the research unit led by Mizuno-san. Since my research up to that time had been in the AI field, I was a bit uneasy about changing my theme to quantum, while at the same time feeling hopeful about getting back into computing. Quantum being a new field, I was given a chance to study up on it.
Sato: Our theme this time is technology called “a shuttling qubit method.” In June 2023, Hitachi announced that it had proposed technology for efficiently controlling qubits, the information units in a quantum computer, aimed at practical implementation of a silicon quantum computer, and had confirmed its effectiveness in a simulation. Among approaches to quantum computers, that based on superconductivity was one of the earlier ones; but Hitachi believes the silicon quantum computer approach is superior for achieving large scale. The shuttling qubit method is one technological approach to large-scale integration in a silicon quantum computer. We contributed to realizing this technology on the software side.
Miyamoto: It was our team that came up with the software technology for efficient control of a quantum computer with large-scale integration. On the hardware side, Hitachi had earlier developed the two-dimensional silicon qubit array, enabling silicon qubits to be integrated by arranging them in a lattice shape. The question remained, however, as to whether the overall computing could be performed efficiently when such a structure is scaled; and the reality is that the control is not easy, as errors creep in and there is crosstalk among the qubits. The shuttling qubit method is looked to as a technology able to control the qubits to minimize errors and crosstalk.
A "shuttling qubit" method for efficient control of qubits
Shuttling is a technology for moving qubits to “lattice cells” (quantum dots) in different arrays. Being able to move qubits while maintaining their quantum state would bring about new possibilities for solving the issues with scaling. One of these issues is that large-scale integration makes the control circuitry more complex. To control qubits, circuits for operations and readout need to be connected to all qubits; but if it becomes possible to move the qubits freely by shuttling them, the operation and readout circuitry will be needed only in specific “lattice cells,” greatly simplifying the wiring structure. One more issue is crosstalk (error) resulting from the influence of qubits on each other. This, too, can be prevented by shuttling, moving aside qubits in adjacent “lattice cells.” To realize this kind of shuttling in arrays on a silicon quantum computer, we are developing a scheduling technology, which determines the sequence in which qubits should be moved.
A key to the development is the “procedure for bringing qubits to operation areas”
Miyamoto: Shuttling is itself an ordinary concept. In initial considerations of the basic approach to a silicon quantum computer, priority was given to scaling by blanketing with qubits. In the course of our studies, when we came up with the idea of using technology for freely moving qubits by shuttling as a computing method, I believe this was a new concept. The thinking was that although spatial efficiency would drop, performing operations while moving qubits would lead to solving the problems.
Development of the shuttling qubit method started out from experimental demonstrations by hardware developers of the principle that electrons can in fact move inside an array. The hints gained from them were highly important, as what we could do as software developers alone was not enough. After thorough discussions, on that basis we developed the software for controlling qubit shuttling, helping to open the way to feasibility of the longed-for quantum computing.
Sato: There are many kinds of constraints on the device side for implementing shuttling. It was necessary to move qubits to operation areas by shuttling while dealing with these constraints. The technology for generating this procedure was central to our accomplishment. Reducing the number of operation circuits, for example, gives rise to the problem that the same operation takes place on qubits in the same row or column of an array. To fix this problem, the operation can be limited to the target qubit by shuffling out only that qubit from the row or column. We were able to develop the software through a process of trial and error while repeatedly running simulations.
Miyamoto: Hitachi is promoting “top down” development premised on scaling. This is why, even at the present time when this scaling has not yet been achieved, R&D is being conducted on software technology to support large-scale integration, like the shuttling qubit method. A major goal is development of a quantum OS, needed for a large-scale quantum computer; and one of these technologies is the shuttling qubit.
The quantum OS, holding the key to quantum computer realization
Sato: In the quantum world, along with development of quantum computer devices, progress is being made in proposing applications making use of quantum algorithms. It would be premature, however, to say that sufficient studies have been conducted on the software development environment or system software for connecting these aspects. Today, the software connecting quantum computer devices and quantum applications is the quantum OS. Its role does not stop with the OS or other system software in a conventional computer, as the concept extends to include something in the nature of middleware.
Miyamoto: Many roles are demanded of a quantum OS. While taking into account the requirements on the hardware side, including quantum computer devices and control units with their many constraints, and considering what kinds of instructions and control signals are to be input from the software side for actually running applications, we are thinking of incorporating in the quantum OS functions like those of a compiler, for automatically translating these instructions and signals.
To this end, we software developers need to actually see and understand how silicon quantum computer researchers are performing experimentation using what devices. So we make frequent visits to labs, looking over the shoulders of those running silicon quantum computer experiments, then go back to developing and thinking. While the fields responsible for hardware and software differ from each other, in this project, the development is being carried out in close unison toward practical implementation of the silicon quantum computer “as a computer.”
Sato: At our project meetings, a wide range of diverse topics are taken up, from the application layer to qubit devices. When topics far afield come up, there are sometimes things we don’t understand at first; but even from these we may derive important hints. Solutions to problems are discovered not by sharing our concerns only with researchers in similar fields as our own, but by bringing together those with many different backgrounds and drawing on their strengths. This is Hitachi’s strength and also what makes the project so interesting. In an environment where communication comes easily, I simply do not feel any walls between researchers.
Miyamoto: It would be a good thing if the quantum computer software technology we are developing would show its presence also in the Noisy Intermediate-Scale Quantum (NISQ) device (a quantum computer with no quantum error correction function, expected to be realized in the near future), on which development is currently under way. If and when the silicon quantum computer is unveiled, the quantum OS being made by Sato-san and me will be a front-end tool. As the face visible to users, the quantum OS will, I believe, be of similar importance as the silicon quantum computer hardware.
Sato: In the 2030s targeted by the NISQ, there are hopes for social implementation, as part of a system, of a quantum computer that actually runs. Devices and applications are also essential to that end, but the fact is, these will not run without a quantum OS connecting the two sides. If one of these does not run, it will not be possible to provide value to society. We are proceeding with our research while feeling deep down inside the importance of the quantum OS.
Atsushi MIYAMOTO
Chief Researcher
Center for Exploratory Research,
Center for Exploratory Research,
In my student days, learning about the fundamentals leading to a quantum computer
As I noted earlier, the book Computer Organization and Design by David A. Patterson and John L. Hennessy, which I read in a Japanese translation by Mitsuaki Narita (Nikkei BP), is one work that taught me the appeal of computers. Passed to me in a seminar exercise, and used also in classes, it consists of two volumes and is a famous work that encompasses the fundamentals of computer architecture. For science and tech people not all that familiar with computers, it should seem fresh and easy to understand; and it is also ideal as an introductory work for students wishing to learn about computers. While it is not a book about quantum computers, you can learn about the core concepts; and even now it is a basic work I myself consider worth going back to.
Naoto SATO
Unit Manager
DX Engineering Research Department,
Service Systems Innovation Center,
Center for Digital Services,
Research and Development Group
A drill series that made me think about the source of ideas
This is a book I bought for my child, but the Unko Drill series (Bunkyosha) is quite something. It was amazing to see our child, who won’t even look at an ordinary drill book, get so passionately involved in these drills. I have to hand it to the person who came up with the idea of this series. It even got me thinking if there was some way I could make use of it in my work, or discover some hints from it.
