The fastest supercomputer in the world, China’s Tianhe-2, occupies nearly 8,000 square-feet of space and can perform operations at a speed of 34 petaflops – that’s 34 trillion operations per second. Governments, universities, and large companies need that kind of computing power for modelling simulations that have thousands of independent variables, such as air traffic control, molecular modeling, and cryptanalysis.
Tianhe-2, like all supercomputers, relies on traditional digital design in which each bit represents either a 0 or a 1. But quantum computers, which were first proposed in the 1980s, have atom-sized bits that can represent 0, 1, or a “superposition” of both 0 and 1 at the same time.
Three quantum bits, or qubits, can represent eight values simultaneously, which means quantum computers could be much faster than traditional computers at solving certain kinds of problems.
Researchers have been working on quantum computing for decades, but have only started building physical quantum-computing components in the last few years – and, this week, the US government announced that it will support this work with a multi-year funding grant. IARPA, the research arm of the 17-member United States Intelligence Community, awarded a multi-year grant to IBM to continue its quantum-computing research.
The main challenge facing researchers is the fragility of qubits. Qubits have to be shielded from heat and electromagnetic interference and cooled to near absolute zero (-459 degree F), or else they’ll return errors. IBM’s most powerful quantum computer so far contains only eight qubits, reports Quartz’s Mike Murphy. But the company has made some important breakthroughs recently, including embedding qubits on a computer chip in a lattice formation to more easily detect quantum errors.
IBM plans to encode a number of imperfect physical qubits into a “logical qubit” that can perform quantum computations reliably and without errors. This logical qubit might then serve as the foundation for future quantum computers, which would have many qubits working together.
Quantum computers promise to be particularly good at chewing through optimization challenges. Think of the traveling salesperson problem, which applies to fields such as logistics, microchip manufacture, and DNA sequencing. Given a number of cities and the distances between them, it asks, what is the most efficient route that visits each city exactly once? The problem is simple to solve at first, but as the number of cities increases, the number of routes between them increases exponentially, to the point that the sheer number of possibilities overwhelms traditional computing resources. A quantum computer could solve in one second an optimization problem that would take a single-core conventional computer 10,000 years to figure out, according to Hartmut Neven, director of engineering at Google.
Google made quantum-computing headlines this week when a quantum computer it developed alongside NASA and computing firm D-Wave vastly outperformed a traditional computer in an optimization problem. Google’s results haven’t been peer-reviewed yet, but they suggest that as quantum computers continue to mature, they could open up entirely new ways of dealing with “big problems” ranging from genomics to climate research.