|The cryogenically cooled D-Wave Two chip provides 512 qubits capacity, considerably larger than the company's first commercial quantum computer chip, the 128-qubit D-Wave One. (Photo courtesy of D-Wave Systems)
Computers have come a long way in recent decades, driven by the power available in microprocessors. Intel co-founder Gordon E. Moore at one time (in a 1965 technical paper) proposed that the number of components in an integrated circuit (IC) or microprocessor would double each two years as IC technology developed and semiconductor processes continued to shrink chip dimensions. It was another Intel executive, David House, who is credited with shrinking this "doubling period" to a more realistic 18 months for recent years. In any case, the power of computers has typically grown as the available power of microprocessors has increased, providing a steady increase in computing power year by year. For some computer users, however, this processing growth rate is insufficient for their demanding requirements, and they have sought alternative means of assembling a computer, such as the quantum computer.
Conventional computers use the many transistors within a microprocessor to perform calculations based on the on and off states of the transistors to represent digital ones and zeros, respectively, which form digital bits. Based on the increasing number of transistors in microprocessors, Moore's Law is expected to reach a termination point somewhere between 2020 and 2030, as the circuits on microprocessors begin to reach an atomic scale. At that point, if not sooner, computing is expected to make a transition from its current binary format to an approach known as quantum computing, in which atoms and molecules, rather than transistors, are used for the memory and processing in a computer. Such an approach offers the potential to perform calculations and run routines at rates considerably faster than silicon-transistor-based computers. The quantum computing approach was first theorized by physicist Paul Benioff at Argonne National Laboratory (www.ani.gov) in 1981.
For those who would rather not wait until 2020, a Canadian company, D-Wave Systems (www.dwavesys.com) was started in 1999 with the goal of creating practical quantum computers and the firm has already delivered several systems to key customers. Rather than use microprocessors as in standard binary computers, D-Wave Systems uses superconducting Josephson junctions in its chips. The computing chips consist of loops of niobium metal containing the Josephson junctions, which are essentially two superconductors separated by an insulating material. The superconductors require cryogenic temperatures for proper operation. They are maintained at temperatures close to absolute zero using liquid-helium (boiling point 4°K) cooling within a highly shielded enclosure to prevent RFI or EMI interference.
Directional Current Flow
When cooled to these temperatures, small electrical currents flow around the niobium loops, with the direction of flow representing the different states of a quantum computer bit or qubit: counterclockwise flow is a 0 bit, clockwise current flow is a 1 bit, and current flowing in both directions can be used to represent both a 0 and 1 simultaneously. D-Wave Systems employs these devices somewhat differently than most researchers involved with quantum computing, who are trying to build quantum logic gates from the devices as building blocks for universal quantum computers. Rather, the team at D-Wave Systems uses an approach called adiabatic quantum computing to process qubits with the novel chips. The approach involves initializing a group of qubits to their lowest energy state and then powering on interactions between the qubits adiabatically, creating a quantum algorithm. The qubits will drift to a new lowest energy state and the qubits can be read for the computational results of the algorithm. In theory, a quantum computer with a given number of qubits can store and process exponentially more information than a conventional binary computer with the same number of bits.
The D-Wave Systems' quantum computers operate with several components, including qubits, couplers, which connect the qubits together, and programmable magnetic memory, which allows each qubit and coupler to be programmed so that problems can be solved. Some noteworthy companies, including IBM (www.ibm.com) and Hewlett-Packard Co. (www. hp.com), have made considerable efforts at assembling a working quantum computer without success. But little D-Wave Systems, with backing from significant investment firms, including Goldman Sachs (www.goldmansachs.com), and collaboration from major organizations such as Google (www.google.com) and NASA (www.nasa.gov), was able to do it.
Not for Your Accountant
D-Wave Systems' quantum computers are not designed for general-purpose use, but for solving specific problems, such as found in the mathematics of science and business. As Vern Brownell, CEO of D-Wave Systems, explains: "Frankly, we don't want thousands of customers, we want a handful of really deep collaborative customers to work on how they can harness this kind of technology, so in the initial phase it's relatively low volume, low number of customers that we are selective about. On that list are the DoD and the intelligence community."
D-Wave sold its first quantum computer, the D-Wave One™, to defense and advanced technology company Lockheed Martin (www.lockheedmartin.com) in 2011, allowing it to bill itself as "the world's first commercial quantum computing company." Lockheed Martin has applied the powerful computer not only to military/electronic-warfare (EW) issues, but to solving challenges involving millions of lines of software code in developing new software programs. In partnership with the University of Southern California (www.usc.edu), Lockheed Martin has established a home for the quantum computer, the USC-Lockheed Martin Quantum Computation Center (QCC). Lockheed Martin upgraded to a D-Wave Two earlier this year. The computer's 512-qubit processing chip is maintained within a cryogenic system capable of absolute-zero temperature within a 10m2 shielded enclosure.
The firm recently sold another of its D-Wave Two™ quantum computers to Google for use at a NASA laboratory for tackling problems requiring a certain amount of creativity. The D-Wave Two is based on a processing chip with 512 qubits capability (see figure), considerably larger than the company's first quantum computer, the 128-qubit D-Wave One. The computer system is being installed at the Quantum Artificial Intelligence Laboratory at NASA Ames Research Center (Moffett Field, CA). The laboratory is collaboration among NASA, Google, and the Universities Space Research Association (www.usra.edu), a granting organization devoted to space exploration and technology. The USRA, which will make the quantum computer system available to other US academic institutions, and research team hope to solve problems related to machine learning, speech recognition, and other functions involved in space research.
Although the number of critics are many for D-Wave Systems and its approach to quantum computing, several investigators who have evaluated the performance of the D-Wave One and D-Wave Two computers have been impressed by their findings. For example, Catherine McGeoch, Beitzel Professor in Technology and Society (Computer Science) at Amherst University (www.amherst.edu), has developed tests to evaluate the processing power of D-Wave Systems' quantum computers in comparison with conventional binary computers. With 25 years of experience setting up experiments to test different aspects of computing speed for different systems, she is one of the founders of "experimental algorithmics," an unusual avenue of computer science. She spent a month at D-Wave Systems to evaluate the first iteration of the company's quantum computing system, finding that it was superior to binary computers in many aspects of operation.