Image Courtesy: IBM
IBM’s latest quantum computing processor has completed a complex computational task in just 19 seconds, a feat researchers estimate could require a leading conventional supercomputer around 110 years to reproduce. The result offers a new demonstration of quantum computing’s potential to outperform classical systems on specific problems.
The experiment used IBM’s 120-qubit Nighthawk r2 processor, accessed through the company’s cloud platform. Researchers selected 61 qubits and ran increasingly complex quantum circuits, ultimately generating one million samples in 19 seconds.
The team, led by theoretical condensed matter physicist Tigran Sedrakyan of BlueQubit, used a benchmark known as random circuit sampling. This involves applying sequences of operations to quantum bits, or qubits, to create complex quantum states. Measuring these states repeatedly produces strings of binary digits that researchers can analyze.
While quantum processors can generate these samples directly, reproducing the same results through conventional computing becomes increasingly difficult as the number of qubits and operations grows.
For the experiment, the researchers ran circuits containing 36 cycles and 918 two-qubit gates. They then used tensor-network contraction, a mathematical technique for estimating the computational resources needed to simulate quantum circuits on classical computers, to assess the challenge.
Their calculations suggested that reproducing the million samples would require approximately 1.2×1027 computational operations. Based on estimates of the sustained performance of Frontier, a leading supercomputer, the team calculated that the task could take around 110 years.
However, the estimate comes with important qualifications. It does not establish an absolute limit on classical computing performance. More efficient algorithms or improved simulation techniques could significantly reduce the time required, as researchers have demonstrated in previous quantum computing experiments.
The latest result also builds on Google’s 2019 quantum computing announcement, when its Sycamore processor performed a random circuit sampling experiment that was initially considered beyond the practical capabilities of classical supercomputers.
A key distinction in the new research is accessibility. The team used commercially available quantum hardware through IBM’s standard cloud workflow rather than specialized laboratory equipment. The researchers also released their circuits, samples, and analysis code to allow others to examine and reproduce the findings.
The experiment marks another milestone in the race to demonstrate practical quantum advantage, although it does not mean quantum computers can outperform conventional machines across everyday computing tasks.
