A new benchmark for comparing quantum computers: What does QUOPS show?
Tested on Google, IBM, and Quantinuum systems, QUOPS has shown that speed and problem-solving capacity in quantum computers are not the same thing.
12punto
Comparing quantum computers solely by looking at the number of qubits or processing speed can often be misleading due to their different architectures. Researchers led by Dr. Timothy Proctor from Sandia National Laboratories have developed a new benchmarking system called QUOPS to provide a common measurement framework for this problem.
QUOPS, which stands for "Quantum Universal Operations Performance System," aims to evaluate both the size of the problem a quantum computer can handle and the speed at which it can do so. This is intended to allow for a clearer comparison of systems that use different types of qubits, connectivity structures, and processing methods.
In the study, which has not yet undergone peer review, QUOPS was tested on advanced quantum systems from Google, IBM, and Quantinuum. The resulting picture showed that the fastest-running system does not necessarily mean the system capable of solving larger quantum problems in every case.
SPEED AND CAPACITY DO NOT PROGRESS IN THE SAME DIRECTION
Google's Willow system reached a speed of 2.0 × 10^7 QUOPS per second and received a score of 216 QUOPS. IBM Boston was measured at a speed of 3.1 × 10^5 QUOPS/s and a value of 204 QUOPS.
On the Quantinuum side, higher problem capacity was notable despite lower speeds. The H2-1 model reached a score of 1320 QUOPS at a speed of 353 QUOPS/s, while the Helios-1 model reached a value of 1504 QUOPS at a speed of 303 QUOPS/s. These results reveal that quantum performance cannot be explained by a single speed indicator.
According to the researchers, QUOPS can be used not only to rank existing devices but also to monitor how close quantum computers are getting to real-world problems. Tasks considered difficult to achieve with classical computers, such as breaking RSA-2048 encryption or solving the reaction mechanism of the FeMo-co enzyme, require capacity far beyond that of current systems.
The study states that for such goals, the capacity of quantum computers needs to be increased by approximately five times and performance needs to reach the level of hundreds of millions of QUOPS. Therefore, the true significance of this new benchmark lies in its ability to monitor different quantum architectures on the same ground, providing a clearer view of how close the technology is to the threshold of practical use.