Programmers are being given a free opportunity to break the ice on a new era of quantum computing. Both D-Wave Systems and IBM are offering free access to their superconducting quantum computers in the cloud, along with access to their continually evolving programming tools.
However, analysts say the software tools to fully utilize the capabilities of quantum computers—by ordinary programmers instead of quantum physicists—will take years to develop, perhaps a decade.
"It will take years to fully develop the quantum computer infrastructure that will challenge the supercomputers we use today," said Robert Sorensen, research vice president of IDC's high-performance computing (HPC) group. "D-Wave and IBM's Quantum Experience [IBM Q] are basically ways of getting programmers to start thinking about how to write software for quantum computers. And since access is free, it offers a golden opportunity to programmers to start writing quantum algorithms and, moreover, define what are quantum computing's grand challenges."
What is quantum computing?
Quantum computers differ from digital electronic computers in a number of ways. For example, standard computing requires data to be encoded into binary digits, or bits, which will always exist in one of two states, 0 or 1; quantum computing utilizes quantum bits (qubits) which can valued as 0, 1, or as superpositions of those states.
Quantum computing could revolutionize computing by solving most NP-Complete (non-deterministic polynomial-time) problems, many of which are impossible to solve on a conventional computer. Quantum computers could also solve many problems conventional supercomputers can but in a fraction of the time, making them perfect for tough problems such as cracking conventional encryptions.
Likewise, NP-Hard problems such as the traveling salesperson problem (requiring the calculation of an optimal route between more than a dozen cities) are practically unsolvable because of the 1012 possible routes that would each have to be calculated and compared by a conventional computer. Superposition, on the other hand, can calculate the length of all possible routes simultaneously, and quickly identify the optimal one.
No wonder so many organizations are in pursuit of "solving the quantum computer problem," including D-Wave, IBM, Intel, Google, Samsung, Microsoft, Quantum Circuits, Rigetti Quantum Computing, the U.S. National Institute of Standards and Technology (NIST), the European Union (EU), Imec (Belgium), Rikken Research (Japan), and many other independent research groups.
D-Wave launched in 1999. The company sold its first quantum computing system to Lockheed Martin Corp. in 2011 (the system is installed at the USC-Lockheed Martin Quantum Computing Center at the USC Information Sciences Institute in Los Angeles). Since then, D-Wave has sold a quantum system to a consortium including the U.S. National Aeronautics and Space Administration (NASA), Google, and the Universities Space Research Association (USRA) (which installed the system at the NASA Ames Research Center in Mountain View, CA); sold another to Los Alamos National Laboratory (LANL) in Los Alamos, NM, and another to Kirkland, WA-based cyberdefense contractor Temporal Defense Systems.
"We also have innumerable customers taking advantage of our cloud offering," said Vern Brownell, president and chief executive officer of D-Wave. Organizations can use D-Wave's quantum annealer free via the cloud. "We also donate time to our software development partners, such as 1-Qubit. D-Wave also selects universities who, along with 1-Qubit, are developing a whole range of open-source tools for buyers of our hardware and licensees of our cloud offerings."
Instead of solving only annealing problems as with D-Wave's devices, IBM, Microsoft, and most of the others have defined various "universal" architectures that can solve most problems conventional computers can solve, plus many they cannot. IBM is using the so-called "gate" architecture, similar to digital computers, in which you first make many individual gates on a chip (such as NAND gates), then interconnect them with metallization layers on top of the same chip, to create higher-level devices (such as multipliers).
"IBM has taken the most visionary approach by going after a 'universal quantum computer" [sometimes called a quantum Turing machine] because it can solve virtually any computing problem, no matter the difficulty," said IDC's Sorensen. "IBM's architecture uses superconductivity like D-Wave, but the full power of IBM's universal quantum computer will only require 50 qubits, but their complexity will vastly outperform the 1,097-qubits D-Wave sells in its systems today."
Sorenson adds, "The advantage of a universal quantum computer is that it can solve the tasks that supercomputers perform today—but in the blink of an eye, instead of hours or days (or longer). In addition, universal quantum computers can create impossible-to-crack encryption algorithms, as well as solve problems hitherto thought to be impossible to solve."
D-Wave's philosophy, according to Brownell, is to only build quantum hardware that can solve problems today. Its quantum annealer is mainly designed for combinatorial optimization problems with many possible discrete solutions. Each time a job is run, the probability of finding the optimal solution is also calculated. In practice, users run the algorithm over and over, each time increasing the probability the solution is optimal, until they achieve a satisfactory probability percentage for that particular problem.
IBM and D-Wave are the only companies offering a working quantum computer for free experimentation online today. IBM says it has only 5 qubits today that are only partially error-corrected (to be upgraded to 20 qubits by 2018). D-Wave acknowledges its 1,097-qubit quantum annealer is not a "universal" quantum computer.
Another promising quantum computing model, called a topological quantum computer requires less error correction but is dependent on the non-abelian anyon, a type of quasiparticle that has not yet conclusively been proven to exist.
Almost all quantum computer prototypes today use superconducting materials cooled to just a few milli-degrees above absolute zero, so thermal fluctuations are minimized and the qubit itself is easier to support in the state of superposition.
"Our approach is with superconducting qubits, which some other groups are pursuing as well, and we'll be selling exclusively IBM Q systems," said Jerry Chow, manager for IBM's experimental quantum computing group. "Our superconducting qubits leverage standard semiconductor fabrication processes, which we think is the most promising route to scale. However, there is still much more engineering and science work to be done before it gets to a level of mass production."
Intel believes qubit functionality can be achieved using standard silicon processes. It may also require a new generation of superconducting sensors, but silicon chips have manufactured with billions of transistors already, making an all-silicon solution at least feasible, according to Intel.
In a recent blog post, Intel chief executive officer Brian Krzanich observed, "Quantum computing is one of the more promising areas of long-term research we've been exploring in our labs, with some of the smartest engineers in the world. We believe it has the potential to augment the capabilities of tomorrow's high performance computers." Krzanich said Intel is collaborating "with the Delft University of Technology (TU Delft) and TNO, the Dutch Organization for Applied Research. TU Delft has been working on the science behind quantum computing for many years and has great vision into the challenges."
R. Colin Johnson is a Kyoto Prize Fellow who has worked as a technology journalist for two decades.