D-Wave Embraces Gate-Based Quantum Computing; Charts Path Forward

By John Russell

October 21, 2021

Earlier this month D-Wave Systems, the quantum computing pioneer that has long championed quantum annealing-based quantum computing (and sometimes taken heat for that approach), announced it was expanding into gate-based quantum computing.

Surprised? Perhaps we shouldn’t be. Spun out of the University of British Columbia in 1999, D-Wave initially targeted gate-based quantum computing and discovered how hard it would be to develop. The company strategy morphed early on.

Mark Johnson, D-Wave

“I joined in 2005 when the company was first transitioning from a gate-model focus to quantum annealing focus,” recalled Mark Johnson, now vice president of quantum technologies and systems products. “There was still this picture that we wanted to find the most direct path to providing valuable quantum applications and we felt that quantum annealing was the was the way to do that. We felt the gate model was maybe 20 years away.”

That was then. Today, buoyed by its experience and success in building semiconductor-based superconducting qubits for annealing systems – notably its systems engineering and on-chip control circuitry expertise, according to Johnson – D-Wave seeks to enter the gate-based market. At its recent annual user meeting, Qubits (virtual), the company spelled out its plans under the banner of Clarity Roadmap.

Much of the roadmap is a continuation of D-Wave’s quantum annealing systems. While makers of gate-based quantum computers struggle to get to 100 qubits, D-Wave has a 5000-qubit system, with 15-way qubit interconnect technology, and is planning a 7000-qubit system. However, quantum annealing and gate-based systems are very different beasts. So-called universal gate-based system quantum computers of the kind being pursued by IBM, Rigetti, Google and others are more flexible and can handle wide range of applications They are the end game.

Quantum annealing systems, such as D-Wave’s Advantage system, have been shown they can solve select optimization problems very well, and in a more direct, less-complicated way than gate-based systems. D-Wave is hardly walking away from quantum annealing computing.

“There’s a growing body of theoretical work now highlighting the fact that quantum annealing is likely always going to be the best approach towards optimization and that’s not what I think we all thought five or ten years ago,” said Johnson, “That’s significant to us and we’ve seen that with a broad range of actually useful customer applications.” In this context, useful customer applications are really more proto-applications (logistics, biopharma, etc.) and no one is using them in production. Still, Johnson believes some apps will be in production in the 1-to-2-year timeframe.

Most of the news around D-Wave’s quantum annealing branch was covered in its press release, which HPCwire posted (D-Wave Details Product Expansion and Cross-Platform Roadmap.) Johnson’s briefing of HPCwire focused on D-Wave move into gate-based quantum computing: why do it now; considerations that drove sticking with superconducting qubit technology and not something else; and a walk-through D-Wave’s five-phase roadmap to develop gate-based systems.

“Let me say this. What we’re proposing to do is going to be hard, right, it’s going to be very hard. There are a lot of very strong competitors in that space. So it’s going to be challenging, but we think we think that our approach is really the strongest,” said Johnson.

So Why Now?

D-Wave contends its work building larger many-thousand-qubit quantum annealing systems has uniquely prepared it for tackling the scale-up problems bogging down gate-based quantum development. Most observers agree that cramming thousands (or millions) of cables into icy-cold dilution refrigerator to control qubits is impractical long-term, and there are many efforts (Intel and Microsoft are just two) to develop cryo-controller chips.

“It looks like we’re going to need hundreds of thousands of qubits and I don’t see how we can connect all the qubits to all the room temperature electronics [with cables]. The only practical, scalable way to develop a gate model system is with some kind of integrated fabrication stack. We think of superconducting-based qubits and a multilayer (3D), integrated circuit fabrication stack, that requires also a lot of investment and understanding of materials and understanding of how to improve coherence. This is exactly what we’re doing for quantum annealing,” said Johnson.

“When we started digging into it, there was a lot of overlap. Many of the problems were very similar. A lot of the technology that we had developed around on-chip integrated control circuitry –  for example, the cold superconducting classical circuits necessary for doing things like homogenizing manufacturing variations across multiple qubits that allowed us to scalable IO so you don’t have several wires for every qubit. The ability to do this sort of de-multiplexed, efficient control of many, thousands or tens of thousands of devices with a just few hundred wires. All of that was necessary to integrate and develop quantum annealing on a [large] scale,” said Johnson.

He didn’t detail the techniques that D-Wave has developed for noise mitigation and signal control but D-Wave has fielded much larger quantum annealing-based systems than those of its gate-based rivals whose systems still hover around 50-60 qubits. Quantum annealing qubits and controls are very different from their gate-based system counterparts, but Johnson says the lessons learned by D-Wave are broadly applicable.

Why Not Ion Trap or Photonics?

It’s perhaps not surprising that D-Wave has chosen to bet on semiconductor-based superconducting qubit technology. That’s what it knows. But there are other choices.

“We tried to take an open-minded look at other qubit technologies because there are some benefits there. There’s the trapped ion approach, and that has some benefits that nature’s ions are all the same. But you move the challenge to a different part of the system. There’s not really a way of getting more than 100 ions in one trap. Then you’re going to have to have multiple traps. And how much have we really demonstrated [the ability] of an algorithm that uses multiple traps? We’ve maybe demonstrated entanglement between them but not practical use of two or three in a single algorithm. When we looked at some of that, we saw difficult challenges in terms of scaling up,” he said.

PsiQuantum wafer

“Then photonics has its own set of challenges. Photons are nice and clean and you can hold information in coherence for a long time. But the flip side of that is it’s hard to do gates with them. They need to be able to interact, and you need some common nonlinearities, and you’ve got to now bring in some kind of atomics to do the interactions. Now, you’ve added another layer of complexity. There are good efforts on all of these different all these different qubit modality threads.”

Another factor in deciding to stick with superconducting quits, said Johnson, was looking at how the different qubit modalities perform on specific applications.

“One thing we did to try and understand [the potential of different qubit approaches] is to look at a problem. So, hypothetically, based on what we know now, factoring a 2000-bit number. You go out and look at the literature and what folks are doing on these system evaluations. With the superconducting model, just sort of using numbers that you can see from the Google effort or the IBM effort or what we can achieve with superconducting technology, you’ve got something on the order of quarter of a day [to perform the calculation].

“But if you do the same analysis, and I’m just looking at what’s been published for trapped ions [and] for photonics, you get to something like on the order of 100 days, once you build it out and look at all the overhead that’s necessary,” he said.

No doubt developers of ion trap such as IonQ and developers of photonics quantum computers such as PsiQuantum have a different view and in fairness to them, it seems like performance metrics, both generally and on specific applications, will improve as all of the qubit technologies advance. It will be interesting to monitor what kinds of benchmarks and qubit technology comparisons emerge from the DOE testbeds and the five National QIS Research Centers set up under the U.S. National Quantum Initiative (NQI).

D-Wave’s Five-Phase Roadmap

D-Wave is betting big on semiconductor-based superconducting. Johnson declined to offer any timelines. However, D-Wave has discussed its overall plan.

“We’ve laid out in our roadmap, this five phase approach. 1) There’s sort of developing our gate model qubit, but within the context of a superconducting integrated circuit stack. This is sort of the first phase. 2) The next phase is validating our model for error correction, and integrated control circuitry. This second phase is going to target a single logical error-corrected qubit comprising something like 16 physical qubits.

“3) The third phase is around demonstrating the scalability of our approach to logical qubit interaction. So that’s targeting a 1000-qubit, physical system, but configurable enough for logical qubits. What that basically will tell us is [if] we’re on the right track and whether we addressed adequately the system engineering problems that will allow us to scale up 4) to the fourth phase of [developing] task-specific components. So do we have a magic state factory, sort of the equivalent of an ALU for qubits, and some sort of memory, etc.? 5) That sets the stage for the first pass to the general purpose QPU, which initially would comprise an error-corrected more generic approach. You can then imagine taking those task-specific modules, and there may be different optimizations for an application, depending on the application.”

Johnson repeated a couple of times during the interview D-Wave’s desire to be a full stack quantum system provider. Just what that means is not entirely clear, particularly given the rapid growth of a varied QIS ecosystem. There are lots of folks trying to build tools that hide underlying quantum technology complexity and make life easier for developers. At this point, it didn’t sound like tapping into that growing community was a goal for D-Wave.

That said, D-Wave would eventually have a portfolio spanning quantum annealing, gate-based systems, and hybrid classical-quantum solvers. Knitting all of those together under a single tool set will be challenging and, at least for now, a go-it-(mostly)-alone approach may be the most efficient.

Of course, the broader question facing the QIS community, including D-Wave, is when will practical uses of quantum computingmove into production environments? There are a ton of “proto-type” efforts now, but nothing in production.

A few people still think that actualization of quantum computing is years away. Perhaps. But a quantum computing boomlet is definitely underway. Government funding. Competing qubit technologies. Advancing quantum networking approaches. Expanding ecosystem (software, hardware, consultants, portals). Industry engagements – indeed D-Wave has one of the largest commercial engagement programs in the industry. Given all of these moving parts, fueled at least in part by geopolitical rivalries, there’s a sense that something concrete and productive will emerge and that today’s quantum bloom won’t repeat the AI crash and long winter that preceded today’s AI boom.

Stay tuned.

Feature Image: D-Wave’s Advantage quantum processor chip

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's latest weapon in the AI battle with GPU maker Nvidia and clou Read more…

ISC 2024 Student Cluster Competition

May 16, 2024

The 2024 ISC 2024 competition welcomed 19 virtual (remote) and eight in-person teams. The in-person teams participated in the conference venue and, while the virtual teams competed using the Bridges-2 supercomputers at t Read more…

Grace Hopper Gets Busy with Science 

May 16, 2024

Nvidia’s new Grace Hopper Superchip (GH200) processor has landed in nine new worldwide systems. The GH200 is a recently announced chip from Nvidia that eliminates the PCI bus from the CPU/GPU communications pathway.  Read more…

Europe’s Race towards Quantum-HPC Integration and Quantum Advantage

May 16, 2024

What an interesting panel, Quantum Advantage — Where are We and What is Needed? While the panelists looked slightly weary — their’s was, after all, one of the last panels at ISC 2024 — the discussion was fascinat Read more…

The Future of AI in Science

May 15, 2024

AI is one of the most transformative and valuable scientific tools ever developed. By harnessing vast amounts of data and computational power, AI systems can uncover patterns, generate insights, and make predictions that Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top500 list of the fastest supercomputers in the world. At s Read more…

Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l Read more…

Europe’s Race towards Quantum-HPC Integration and Quantum Advantage

May 16, 2024

What an interesting panel, Quantum Advantage — Where are We and What is Needed? While the panelists looked slightly weary — their’s was, after all, one of Read more…

The Future of AI in Science

May 15, 2024

AI is one of the most transformative and valuable scientific tools ever developed. By harnessing vast amounts of data and computational power, AI systems can un Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

ISC 2024 Keynote: High-precision Computing Will Be a Foundation for AI Models

May 15, 2024

Some scientific computing applications cannot sacrifice accuracy and will always require high-precision computing. Therefore, conventional high-performance c Read more…

Shutterstock 493860193

Linux Foundation Announces the Launch of the High-Performance Software Foundation

May 14, 2024

The Linux Foundation, the nonprofit organization enabling mass innovation through open source, is excited to announce the launch of the High-Performance Softw Read more…

ISC 2024: Hyperion Research Predicts HPC Market Rebound after Flat 2023

May 13, 2024

First, the top line: the overall HPC market was flat in 2023 at roughly $37 billion, bogged down by supply chain issues and slowed acceptance of some larger sys Read more…

Top 500: Aurora Breaks into Exascale, but Can’t Get to the Frontier of HPC

May 13, 2024

The 63rd installment of the TOP500 list is available today in coordination with the kickoff of ISC 2024 in Hamburg, Germany. Once again, the Frontier system at Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Leading Solution Providers

Contributors

Eyes on the Quantum Prize – D-Wave Says its Time is Now

January 30, 2024

Early quantum computing pioneer D-Wave again asserted – that at least for D-Wave – the commercial quantum era has begun. Speaking at its first in-person Ana Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

The GenAI Datacenter Squeeze Is Here

February 1, 2024

The immediate effect of the GenAI GPU Squeeze was to reduce availability, either direct purchase or cloud access, increase cost, and push demand through the roof. A secondary issue has been developing over the last several years. Even though your organization secured several racks... Read more…

Intel Plans Falcon Shores 2 GPU Supercomputing Chip for 2026  

August 8, 2023

Intel is planning to onboard a new version of the Falcon Shores chip in 2026, which is code-named Falcon Shores 2. The new product was announced by CEO Pat Gel Read more…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh Read more…

GenAI Having Major Impact on Data Culture, Survey Says

February 21, 2024

While 2023 was the year of GenAI, the adoption rates for GenAI did not match expectations. Most organizations are continuing to invest in GenAI but are yet to Read more…

How the Chip Industry is Helping a Battery Company

May 8, 2024

Chip companies, once seen as engineering pure plays, are now at the center of geopolitical intrigue. Chip manufacturing firms, especially TSMC and Intel, have b Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire