INTELLIGENT TECHNOLOGY

HIGH-PERFORMANCE COMPUTING

The initiative is being led by three senior figures from the Quantum High- Performance Computing Collaboration Platform Division at the RIKEN Center for Computational Science in Kobe.

The lead researchers are Division Director Mitsuhisa Sato and Deputy Director Yusuke Kodama, both of whom contributed to the development of one of the world’ s fastest supercomputers, Fugaku. They are joined by Deputy Director Tamiya Onodera, who joined the division in April 2025 after working at a quantum research lab in Tokyo for global technology company IBM.

By leveraging quantum superposition, which allows quantum computers to process multiple possibilities simultaneously, quantum systems can handle far more information than conventional computers that process only one data state at a time. As a result, quantum computers have the potential to solve problems that are extremely challenging for conventional machines at high speed.

Today’ s quantum computers use a variety of approaches, including ion traps, superconducting circuits, optics and silicon-based technologies. Despite their growing performance, integration with supercomputers remains essential because quantum and conventional computers excel in fundamentally different areas. while powerful for specific problems, remain experimental and are not designed for routine or data-heavy workloads.

Sato explains that quantum computers can solve problems that supercomputers struggle with but currently require conventional computers for control.“ In the future, as quantum computers improve ten-fold or even a hundred-fold, control and communication will require much more help from supercomputer-level computing,” he said.

Supercomputers also struggle with what Sato describes as a‘ computational explosion’. As the number of binary options increases, the number of possible combinations grows exponentially, demanding enormous processing time. Quantum computers are particularly effective in these scenarios, making them valuable for materials development, drug discovery, artificial intelligence and optimisation tasks.

Currently, leading quantum computers have just surpassed 100 qubits. Unlike conventional bits, which are either 0 or 1, qubits can exist in multiple states at once. However, quantum computers still rely on conventional systems to interpret and execute commands. Sato compares this relationship to a piano, sheet music and a pianist.

As qubit counts rise to 1,000 or even 10,000, supercomputer-level performance will be essential. To address this, the RIKEN-led JHPCquantum project is developing core system software to link quantum computers and supercomputers. The five-year project, launched in November 2023, brings together RIKEN, the University of Tokyo, Osaka University and SoftBank Corporation.

Two quantum systems are currently in use. The ion trap-based Reimei system was introduced in February 2025 at RIKEN’ s Wako campus. IBM’ s superconducting IBM Quantum System Two, known as ibm _ kobe, was introduced in June at the Kobe campus.

According to Onodera, exploring both technologies ensures long-term software compatibility.“ Superconducting quantum computer types will likely achieve more than 10,000 qubits, but no one really knows which technology will reach one million qubits,” he said.

The systems are now being evaluated by 12 user groups.

“ From now on, actual evaluations will be conducted using the programming environment on the system software we have developed,” said Kodama. •

Supercomputers are well suited to general-purpose tasks, large-scale simulations and reliable processing of massive datasets. Quantum computers,

28

INTELLIGENT CIO APAC www. intelligentcio. com