D-Wave will ship a 2000-qubit quantum computer next year
- 21 September, 2016 06:26
Forget PCs and servers: D-Wave Systems is looking into the future with its quantum computer, up to 1,000 times faster than an earlier model.
The company will start shipping a quantum computer with 2,000 qubits, twice the size of its current 1,000-qubit D-Wave 2X. The D-Wave 2X is considered one of the most advanced computers in the world today.
The 2,000-qubit quantum computer will be 500 to 1000 times faster than its predecessor, said Jeremy Hilton, senior vice president of systems at D-Wave.
An even larger quantum computer based on a whole new processor design will come out two to three years after that, Hilton said.
Today's PCs and servers could ultimately be replaced by a quantum computer, which has been researched for decades. Beyond D-Wave, companies like IBM are also building quantum computers.
D-Wave deployed what was considered the first quantum computer in 2011, and its systems have been purchased by Google, Lockheed Martin, NASA, and the Los Alamos National Laboratory.
The ultimate goal is to develop a universal quantum computer, which is being pursued by IBM. There are many ways to build quantum computers, and D-Wave's system -- which is based on the concept of quantum annealing -- is aimed at specific tasks. IBM's using a different model to build quantum computers that can run multiple applications, much like PCs.
A larger processor will boost the computing capability of the upcoming D-Wave quantum system, which could speed up applications like machine learning, engineering and software validation, Hilton said.
D-Wave is still discovering uses for its quantum computer. Los Alamos National Laboratory bought a D-Wave system to study new computer designs and uses. Google is using the computer for machine learning and other applications. D-Wave is also offering remote access to its quantum computer to specific users.
One promising area is machine learning. D-Wave itself isn't a machine learning company, but it is developing models that could improve image classification, generation, and analysis. D-Wave also sees uses of the quantum computer in the health and finance sectors.
In the future, it'll be possible for quantum computers to swiftly recognize a person in images as soon as the images are uploaded. The process should be much faster than with today's servers.
D-Wave is hosting a user meeting next week to encourage software development for its quantum computer. The development can happen using common languages like C++ and using popular machine learning models like TensorFlow and Caffe.
The company wants to make it easier for programmers to transition to a new type of computer by using conventional coding models, Hilton said.
Quantum computers use different hardware than today's PCs but building them has been difficult because qubits, central to the systems, can be unpredictable. PCs typically store data in 1s and 0s, but qubits can be store data in multiple states, which allows quantum computers to do more calculations simultaneously.
Qubits harness the laws of quantum mechanics to achieve various states, but are unstable and can quickly go out of control with the smallest of interruptions or disturbances. The instability could affect applications like genome sequencing, which need sustained processing capabilities for long periods.
The behavior or state of qubits -- whether it'll be a one or zero or another state -- is hard to predict once they start interacting, or "entangling.” However, D-Wave has achieved a level of consistency in stabilizing the logical qubits in its quantum computers. The qubits are loaded in a controlled environment that is cooled and heavily pressurized.
Some critics have argued that D-Wave's technique isn't true quantum computing, while other researchers have validated the method. Either way, D-Wave is providing a powerful computer benefiting its customers.
D-Wave is also going after the universal quantum computing model, Hilton said. But the short-term goal is to discover more use models in areas like artificial intelligence where D-Wave's quantum computer might hold promise.
With all the hype of quantum computers, PCs and servers won't go away anytime soon. More research is needed to build a universal quantum computer. The only competition to D-Wave is IBM, which is offering cloud-based access to a 5-qubit processor for people to experiment with.