(Source: UCSB)
Professor John Martinis' "xmon" junctions may drive the next generation of D-Wave Systems quantum computers

The University of California, Santa Barbara (UCSB) -- home to the Kavli Institute for Theoretical Physics -- is teaming up with Google Inc. (GOOG) to further its commercialization of superconductor-based quantum computers.  Google plans to use the collaboration productize and scale up production of its Canadian partner D-Wave Systems, Inc.'s quantum processor designs.
I. Long Shot
D-Wave seemed a long shot when it launched in 1999.  An offshoot of the University of British Columbia (UBC), the company looked to commercialize what in the 1990s was a relatively controversial and theoretical field -- quantum computing.
Quantum computing is the art of manipulating atomic components such as electrons and exploiting atomic-scale quirks of physics known as "quantum effects" to densely store information and nearly instantaneously derive answers of complex algorithms.  Quantum computing is a two-sided challenge.  First, you need to develop hardware capable of exploiting the desired kinds of quantum manipulation.  Second, you need to program that hardware with a quantum algorithm derived from a more classical algorithm, such as Google's Page Rank search algorithm.
In the 1990s neither the hardware nor the quantum algorithms development was advanced enough to be considered commercially viable.  But D-Wave rolled the dice and launched a startup hoping it could change that on the hardware side.  Working closely with the UBC; the University of Toronto; The National Aeronautics and Space Administration's (NASA) Jet Propulsion Laboratory microdevices lab in Pasadena, California; and various other American and Canadian research institutions, D-Wave managed to achieve the unlikely -- building quantum computing hardware with commercial potential.
In 2007 it performed its first public demo on the 16-qubit (qubit = quantum bit) "Orion" system, which had at its hard a special piece of hardware called a "superconducting adiabatic quantum computer processor".  The quantum processor was more of a coprocessor, in the sense that it was not designed for general purpose computing, but rather ran specialized quantum algorithms loaded onto it by conventional hardware. 

D-Wave System
To perform reliable quantum computation, D-Wave Systems' designs operate at near absolute zero temperatures. [Image Source: ZDNet]

Solutions were generated based on a complex process known as quantum annealing.  More specifically, the D-Wave system operates via quantum entanglement -- a sort of psychic link (metaphorically speaking) between electrons in which they mirror each others' states (more specifically, mirror each others' spins about their respective atomic nuclei).  To achieve reliable quantum results, the system is chilled to a temperature as cold as or colder than outer space.
D-Wave Systems' claim of having the first working quantum computer was a controversial one.  However, skeptics were unable to disprove its claims and supporters such as NASA eventually published evidence supporting their possible validity.

16-Qubit Orion
D-Wave Systems' first publicly unveiled design was the 16-qubit "Orion" processor.

From the start applications to search were apparent.  Along with solving Sudoku puzzles and a scheduling problem, one of the initial three demos in 2007 involved searching for known molecules in a database.  Certain graph search algorithms are NP-complete problems; hence it's impossible in a conventional sense to generate a precise solution in most cases.

D-Wave chips
D-Wave began early commercial production with the pictured "Orion" 16-qubit processor.

The goal is to develop heuristics to give a good approximate, based on the needs.  One of the core goals of D-Wave was to develop commercial hardware to derive approximate solutions to NP-complete problems at rates far faster than is possible with conventional hardware.  D-Wave was building hardware to perform searches on steroids, so to speak.
In May 2011 D-Wave Systems announced the availability of a 128-qubit processor dubbed "Rainier" (codename: Chimera) and supporting hardware.  The full system was dubbed the "D-Wave One" and had a price tag of $10M USD.  This system has a typical operating temperature of 13.8 millikelvin (mK).  By contrast the coldest naturally occurring temperature in space is around 1 K, while the average space temperature is around 2.7 K -- almost 200 times the temperature inside the quantum computer.

D-Wave chip
D-Wave in 2012 released 128-qubit quantum processors.

Despite the relative expense, the system received several purchases.  Clients included Harvard University, Lockheed Martin Corp. (LMT), and Cornell University.  While the system showed questionable results in terms of acceleration, it at least debunked skepticism by providing evidence that the complex hardware was truly functioning as described.

D-Wave One
The D-Wave One quantum computer retailed for $10M USD.

Later that year D-Wave Systems teased at a 512-qubit system code-named Vesuvius.  That system was released last year under the name D-Wave Two.
The D-Wave Two design was a pivotal leap as it finally began to eclipse traditional computers in speed.  In some cases, the D-Wave Two was hundreds of times at fast at performing complex problems when pitting the best quantum algorithm against the most fit classical algorithm on state-of-the-art traditional hardware.
II. Google and D-Wave: Entangled for a Common Cause
Google's involvement with D-Wave dates back over half a decade.  In 2009, one of the first demonstrations of D-Wave's early prototypes involved a Google quantum image search algorithm that found pictures of cats.
In May 2013 Google and D-Wave Systems deepened their partnership, announcing the opening of the Google Quantum Artificial Intelligence Lab.  The lab was cosponsored by NASA and the Universities Space Research Association (USRA).  It showed that in various benchmarks D-Wave One and Two systems could be made to offer an acceleration of anywhere from 3 to 5 orders of magnitude over conventional algorithms and hardware.
Google used its D-Wave Systems boxes to optimize parts of its Android operating system; the world's most used mobile operating system.  While it declined to reveal certain specifics, Google said the optimizations it achieved with the quantum hardware were well beyond what was feasible using its state-of-the-art conventional hardware.  It said that one D-Wave Systems box could do the optimization work, in some cases, of an entire large data center.

This May the lab offered up a crucial validation of D-Wave Systems' technology, using a bleeding edge technique called qubit tunneling spectroscopy.  Using this imaging technology, Google, D-Wave Systems, and various academic partners tested cut-down 2-qubit and 8-qubit D-Wave designs.  They observed coherence in the devices' energy eigenspectrum, strong evidence of quantum entanglement during a key portion of the design's simulated annealing process.

This validation was very important, as some questions had remained whether D-Wave's machine was truly accomplishing quantum annealing.  While quantum physics descriptions provided the best fit to previously observed metrics, classical models were shown in studies to possibly demonstrate somewhat similar behavior.  The new work from Google, et al., by comparison showed relatively unambiguously that quantum mechanisms were at work within D-Wave's mysterious boxes.

III. New Blood

The just announced partnership with UCSB brings aboard one of the world's top experts on superconducting, physicist John Martinis, PhD.  Professor Martinis has won the London Prize -- a top research accolade -- for applications of superconducting to quantum computing.  His expertise lies in quantum control -- designing chemical systems that allow ready manipulation of quantum states -- and quantum information processing -- designing specialized quantum algorithms (software) to parse complex graphed datasets.

Martinis Group
The Martinis Group: Austin Fowler, Rami Barends, Professor John Martinis and Julian Kelly

Google's Director of Engineering Hartmut Neven reports:

With an integrated hardware group the Quantum AI team will now be able to implement and test new designs for quantum optimization and inference processors based on recent theoretical insights as well as our learnings from the D-Wave quantum annealing architecture. We will continue to collaborate with D-Wave scientists and to experiment with the "Vesuvius" machine at NASA Ames which will be upgraded to a 1000 qubit "Washington" processor.

D-Wave Systems revealed last year that its qubit design consists of superconducting loops comprised of niobium (an elemental superconductor) with an insulating layer of aluminum oxide at the junction.  The superconducting loops are know as Josephson junctions, in honor of British physicist Brian David Josephson who won a Nobel Prize for describing in 1962 the behavior behind this kind of circuit.

Superconducting circuits of niobium and aluminum oxide can be grown on a silicon substrate for mass-produceable quantum electronics designs. [Image Source: IEEE Spectrum]

The phase and the charge of the superconductor used have been shown to be key determinates to how long qubits can be maintained, and what levels of entanglement can be achieved during quantum annealing.  Niobium phased out lead as the superconductor used in these junctions [source], which are also known as superconducting tunnel junctions (STJs).
There's no clear replacement to Niobium, which is perhaps the highest temperature elemental superconductor, with a critical superconducting temperature of 9.26 K.  Niobium is relatively abundant and cheap, as far as rare metals go.  Research has suggested that Technetium under high pressure could superconduct at up to 11.2 K [source], but more work needs to be done to examine the feasibility of using such a material in an STJ.

Niobium Crystals
Elemental Niobium in alloy form is blue, and is found in crystalline deposits.
[Image Source: Wikimedia Commons]

In the meantime Professor Martinis can focus his expertise on optimizing the junction's geometry and the deposition techniques to produce it consistently and affordably.

Professor Martinis will likely also work to productize a special kind of Josephson junction geometry -- a cross-shaped junction he calls an "xmon".  His team and their collaborators in April published a paper on near-commercial quality xmons in one of science's most prestigious peer-reviewed research journals, Nature.  Xmons show superior entanglement to other junction geometries, but (as the paper states) are just starting to approach commercial readiness on the production front.

Xmon chip
Professor Martinis has developed an improved Josephson junction, which has a cross shaped geometry.  He calls it an "xmon" qubit. [Image Source: UCSB]

The payoff for Google will in short term largely be realized by using software optimization algorithms that hunt for inefficiencies in a codebase.  In the longer run, Google may be able to sink the prices of these quantum systems low enough to make them usable as a search backend to its images engine, provide a much faster and "smarter" search.

Source: Google Research [blog]

"Intel is investing heavily (think gazillions of dollars and bazillions of engineering man hours) in resources to create an Intel host controllers spec in order to speed time to market of the USB 3.0 technology." -- Intel blogger Nick Knupffer

Latest Blog Posts
Xiaomi Mi 6 Smartphone.
Nenfort Golit - Aug 8, 2017, 6:00 AM

Copyright 2017 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki