Riverlane's flagship QEC breakthrough earns recognition as a significant advance towards fault-tolerant quantum computing.
2024 has been full of milestones in quantum error correction (QEC), but the journey is just beginning. We are still at a tipping point in the QEC Era where we need to reach the one million operations threshold, known as the MegaQuOp, to surpass classical supercomputers and unlock quantum computing's transformative power.
Nonetheless we must celebrate each and every milestone. Because today we are farther along on this journey than I think anyone in the quantum industry thought we would be - even five years ago. That is why for Riverlane, yesterday was a proud day. Our landmark QEC paper, presenting the results of the first comprehensive hardware implementation in quantum computing using the company’s decoder chip, was published in Nature Electronics and recognized as a breakthrough for QEC and the world's first dedicated 'QEC chip' implemented on ASIC and FPGA hardware. By uniquely balancing speed, accuracy, and resource efficiency, it set a new standard for quantum error correction when it was first published on arXiv in September 2023.
Quantum error correction is a set of techniques used to protect the information stored in qubits from errors and decoherence caused by noise. Decoding is one of these techniques. Large-scale quantum computers will generate high volumes of data every second, scaling to 100TB/s, and the system must decode this data faster than it accumulates to prevent errors from building up and rendering calculations useless.
Riverlane’s quantum decoder is the central element of Riverlane’s core product, the ‘Quantum Error Correction Stack’, Deltaflow, responsible for identifying underlying errors in a quantum computer and issuing corrective measures.
"Deltaflow not only decodes fast enough to keep up with a quantum computer but also only uses a tiny resource footprint. This combination significantly impacts the quantum computer where our decoder's high performance, small-area features provide easy integration and low-power consumption of less than ten mW – about the same as a smart watch in sleep mode." said Earl Campbell, VP of Quantum Science at Riverlane.
“I’m excited by Riverlane’s breakthrough result, reported in Nature Electronics, and what it means for the quantum computing industry. Performing quantum error correction is a critical hurdle in realizing practical quantum computers, and Riverlane’s scalable, high-performance quantum decoder is a significant advance. The ability to decoder errors in real-time while maintaining a minimal hardware and energy footprint is exactly the kind of innovation we need to move closer to fault-tolerant quantum computing.” said Stephen Bartlett, Professor of Physics at the University of Sydney.
At its release, our novel Collision Clustering (CC) decoder was the most powerful in the world, and this paper was among the first to demonstrate that QEC could provide a viable route to fault-tolerant quantum computing. Since then, the industry has rapidly embraced QEC, with companies like Google, Quantinuum, and QuEra also making significant QEC advancements.
But there is more work to be done before we get to practical applications. The next big milestone? Using real-time QEC to achieve a 'MegaQuOp' (one million error-free quantum operations). This is where quantum computing's first practical use cases beyond the supercomputing threshold will begin to emerge.
Over the past 20 years, qubit error rates have halved every 12-18 months. They have now got to the breakeven point where QEC starts to work. Adding QEC means we will be able to decrease the error rates by 10,000x reaching 1mn error free operations (MegaQuOp) within the next 2-3 years. In order to build a Mega QuOp quantum computer you need two things:
- 1,000 – 10,000 physical qubits with physical error rates below the QEC breakeven point
- The QEC stack: The (classical) compute needed to process the QEC data in real-time
The team at Google, for example, now have 100 physical qubits at the breakeven point (and will have by next year gone sufficiently below the breakeven point). Their roadmap gets them to (1) in the next 2-3 years and this continues to look credible. The next generation of quantum computers will require (2), something Riverlane plans to address in 2025.
Our recent QEC report explains in more detail how crossing into the million error-free quantum operations (MegaQuOp) regime presents a pivotal moment in quantum computing, when exactly the industry expects to achieve this critical milestone and predictions for what this will look like.
Here Riverlane is leading the charge. Our Quantum Error Correction Stack (Deltaflow ™), the first comprehensive QEC solution for all quantum computers, will help our hardware partners transform their full-stack quantum computers into 'MegaQuop Machines'. With our second-generation QEC Stack (Deltaflow 2) launching this March, we're steadily advancing toward this critical goal.
“Many people don't realize that in a quantum computer, a classical error correction layer is essential to keep the quantum machine on track. For many years, there has been debate as to whether conventional computers can perform this complex task fast enough to keep up. This beautiful work from the Riverlane team settles the question. Fortunately, the conventional circuits can keep up." said Simon Benjamin, Professor of Quantum Technologies at Oxford University.
Yesterday’s publication in Nature Electronic was a proud moment. With QEC implementation now a top priority for every quantum computer and control system maker, we look forward to driving the continued progress of our QEC roadmap through 2025 and beyond and celebrating many more proud moments.