IBM Corp. today announced a new milestone in its quest to develop a large-scale quantum computer, saying that it has doubled its 27-qubit Falcon chip’s quantum volume.

Quantum volume is a metric for measuring quantum computer performance. The Falcon chip boasts a quantum volume of 64, double that of IBM’s previous most powerful system. Only one other system is known to have hit the 64 mark to date: Honeywell International Inc.’s 6-qubit “trapped ion” platform.

IBM has set a goal of at least doubling its hardware’s quantum volume every year and successfully did so in both  2018 and 2019. With today’s milestone, the company can now boast of having maintained the streak for the third year in a row.

IBM achieved the latest performance bump by making enhancements to one of the three 27-qubit Falcon chips it announced earlier this year. The processors are among the newest developed by IBM, which have deployed a total of 28 quantum processors over recent years. The Falcon series is one of multiple chip designs that the company is working on in parallel. 

IBM said it achieved the quantum volume increase through “hardware-aware” optimizations that unlock more performance from the Falcon chip’s circuits. These optimizations include software improvements to Qiskit, IBM’s compiler for running code on its quantum systems. The company’s researchers have also developed a technique for reducing qubit noise, or interference, which is a source of information loss and errors that slow quantum calculations.

One of the most notable details of the announcement is that IBM’s optimizations can be applied to its other chips. In future processors larger than the 27-qubit Falcon, the optimizations could potentially unlock even greater performance jumps. 

IBM is already readying on its next generation of systems. Jay Gambetta, the vice president of quantum at the company, told CNET that his group is developing a new, “greatly” improved version of the 53-qubit Hummingbird processor it detailed at the start of 2020. 

“IBM’s full-stack approach gives an innovative avenue to develop hardware-aware applications, algorithms and circuits” Gambetta added in a statement.

Even with the technical progress made in recent years by IBM and others, there’s still a long way to go before large-scale quantum computing can become a reality. Qubit noise, one of the focus points of IBM’s new optimizations, continues to be a major obstacle. There are also other challenges that researchers are trying to tackle, such as the need to reduce the number of network cables a quantum chip requires to transfer data.

Nonetheless, scientists are already working to come up with ways of applying quantum computing to real-world problems. IBM’s quantum chips play a notable role in that effort. The company has made its chips accessible to researchers via a cloud-based service that counts more than 250,000 registered users worldwide.

Photo: IBM