Quantum computers and conventional supercomputers can serve as powerful tools for analyzing chemical processes. The ongoing collaboration between IBM and Riken, a Japanese scientific institute, is paving the way towards this goal.

Successful chemical analysis often hinges on comprehending how molecules behave during reactions, such as in therapies or industrial catalysts, frequently linked to the quantum state of electrons. Quantum computers can expedite the calculations of these states, yet they remain prone to errors in their current configurations. Traditional supercomputers can catch these discrepancies before they escalate into larger issues.

In a collective statement to New Scientist, Aoki Sei and Mitsui Sato from Riken noted that quantum computers can augment traditional computing capabilities. Currently, they and their team are modeling two distinct iron-sulfur compounds using IBM’s Heron quantum computer in conjunction with Riken’s Fugaku supercomputer.

The researchers divided the computation of the quantum states of the molecules among machines that leverage up to 77 qubits and utilize an algorithm known as SQD. The quantum computer performs the calculations while the supercomputer verifies and corrects errors. For instance, if Heron generates a mathematical representation indicating more electrons than actually present in the molecule, Fugaku discards some of the results, prompting Heron to adjust and retry the computation.

This hybrid approach has not yet surpassed the optimal scenarios achievable by standalone supercomputers, but it competes well against some standard methods, according to Jay Gambetta at IBM, who was not involved in the research. “It’s a matter of comparing calculators,” he remarked.

Recently, this integration is being recognized as the “secret sauce” for addressing the challenges posed by error-prone quantum computers, as articulated by Kenneth Meltz from the Cleveland Clinic in Ohio. His team is employing another IBM quantum computer, paired with a traditional system, to innovate variations of SQD algorithms that model molecules in solutions, offering a more accurate depiction of chemical experiments than past models.

In Meltz’s perspective, advancing the SQD algorithm will enable the combination of quantum and conventional computing to yield substantial benefits over the next year.

“The synergy between quantum and supercomputing is not merely useful; it is an inevitability,” stated Sam Stanwyck from Nvidia. He emphasizes that the future of quantum computing lies in its seamless integration with robust classical and quantum processors from supercomputing centers. Nvidia has already developed a software platform to facilitate such hybrid methodologies.

Aseem Data from Microsoft remarked that his organization is also venturing into groundbreaking possibilities that merge quantum computing, supercomputing, and AI to expedite developments in chemistry and materials science.

Despite these advancements, numerous challenges persist within the quantum computing sector. Markus Reiher from ETH Zurich acknowledged that while the outcomes of the Riken experiments look promising, it remains uncertain if this methodology will become the preferred technique for executing quantum chemical analyses. The precision of the computed results derived from Quantum and Supercomputing partnerships is still undetermined. Additionally, conventional methods for performing such calculations are already established and highly effective.

The potential of integrating quantum computers into computational processes is lauded for enabling the modeling of larger molecules and enhancing processing speed. However, Reiher expresses caution about the scalability of this emerging approach.

According to Gambetta, a new iteration of IBM’s Heron Quantum Computer was launched at Riken in June, boasting reduced error rates compared to its predecessors. He anticipates noteworthy hardware advancements in the near future.

Moreover, researchers have fine-tuned the SQD algorithm to bolster how Heron and Fugaku collaborate in parallel, making the process more efficient. Meltz compares the current status to that of traditional supercomputers from the 1980s, highlighting numerous unresolved issues. Nevertheless, the infusion of new technology promises significant returns.