Despite the fact that quantum computers have not yet proved their practical benefits, they have already generated a number of startups involved in the development of auxiliary technologies. One of these startups is the British company riverlane, which is working on a critically important aspect of quantum calculations – processing the huge amounts of data necessary to correct errors in quantum systems.
The problem that Riverlane solves is the fragility of quantum states. Cubes, the main units of quantum information, tend to lose their condition during operations or simply over time. This limits the amount of calculations that can be performed before the inevitable error. The solution is the use of “logical cubes” that combine many physical cubes for the distribution of quantum information. Additional cubes are used to monitor errors and allow them to correct them.
Founder and CEO of Riverlane, Steve Brierly, spoke about the scale of the problem: when working with a million of physical cubes, it will be necessary to process about 100 terabytes of data per second, which is comparable to the global Netflix stream. Moreover, processing should occur in real time, so as not to slow down quantum calculations.
The company has developed a specialized chip and an algorithm called “Collision Clustering Decoder” to solve this problem. The current version of the chip is able to process data for a logical cube built on the basis of almost 900 physical cubes. At the same time, the chip occupies a tiny area and consumes only 8 millivatts of energy.
Riverlane’s solution is compatible with various types of quantum systems, including superconducting cubes, ion traps, and cold atoms. This makes the technology universal and potentially applicable throughout the industry.
The company has already announced a roadmap on the scaling of its technology. The next generation of the chip will support up to 10,000 physical cubes, and in the future, it is planned to increase the power 10 times every 12-18 months.
Steve Brillie optimistically looks at the future of quantum calculations. It predicts the appearance of the first long-lived logical cubit for the next 12 months, and the achievement of hundreds of logical cubes for 2-3 years. For technology, which many considered forever unattainable, these are very short time.
The success of Riverlane can be a key factor in overcoming one of the main obstacles to creating full-fledged quantum computers that can solve practical problems.