D-Wave calls quantum computers "solvers." And as with PCs today, these solvers are going to segmented into fast, faster, and fastest.
The Canadian company already has a relatively fast quantum computing device, and is going to build a faster quantum computer that can handle more complex questions. This superconducting system is designed to use the gate model that is also the tech behind quantum computers made by IBM and Google.
For the past decade, D-Wave has been selling quantum-annealing systems, which can be used to solve small problems. Gate-model computers, like the one now promised by D-Wave, are theoretically proven to be better for more substantial work, the manufacturer's executives said at their Qubits conference, which was live-streamed.
"This is not an overnight exercise," said Mark Johnson, vice president of quantum technologies and systems at D-Wave. "We're not trying to fool anybody here. There's a lot of required. We do think that some of these technologies ... are going to take time."
D-Wave expects to make this gate-model computer available by 2023 or 2024.
"Our goal over the next few years is to get a small error-corrected [gate-model] system to Leap," Johnson added, referring to its quantum-computers-in-the-cloud service. That error correction part is important, as this is required to deal with finicky qubits.
In quantum computers, information is encoded in quantum bits, or qubits, with values of 0, 1 or superposition with both. This property allows quantum computers to quickly solve large problems that would be infeasible for conventional computers.
Data at the gate
The design of the gate-model computer being targeted by D-Wave is considered by many a direct pathway to a universal quantum computer.
The gate-model system is more complex to build than the company's current quantum annealing technology, which is limited in scope and targeted at solving specific problems. The annealing system delivers a set of possible outcomes after a magnetic field is applied to perform qubit operations.
"We know there are important problems annealing systems can't solve, like differential equations," said Alan Baratz, CEO of D-Wave, during a speech. He highlighted fluid dynamics and quantum chemistry as other examples, and added "they do require the gate model."
- AWS taps up Singapore scientists to overcome hurdles facing quantum computing
- Microsoft's Azure Quantum hits preview: Not so much quantum computing as it is quantum-inspired computing
- 'Quantum computer algorithms are linear algebra, probabilities. This is not something that we do a good job of teaching our kids'
- AWS creates a quantum computing cloud with classical testbed plus rentable qubits
Baratz said D-Wave's quantum gate-model computer and annealing systems will co-exist, serving two separate purposes. In pharmaceuticals, drug discovery will require the massive computing power of gate model, while moving the drug through approval process is more about optimization, which is better suited for annealing, Baratz claimed.
That gate-model computer will pit D-Wave against IBM, which in the past published [PDF] an academic paper arguing that D-Wave's quantum annealing system wasn't true quantum computing.
D-Wave's upcoming system will be based on flux qubits, which are already being used in the annealing system. Big Blue and Google's are based on transmon qubits. The process of building the error-corrected system also involves fabrication of multi-layer qubits and on-chip addressing for I/O management.
The company also announced a 7,000-qubit quantum annealing system, an upgrade from the 5,000-qubit computer announced last year. This new device has a revised qubit design and fabrication stack, Baratz said. ®