Chepko Danil - Fotolia
BOSTON -- IBM has once again doubled the performance of its quantum computing system and plans to keep up the pace, as the new architecture seeks its place in the enterprise.
IBM's new highest Quantum Volume -- the measurement for determining the overall performance of a quantum computer -- shows its 20-qubit processor, IBM Q System One, is now twice as powerful as an earlier version of the system just last year, the company said during the American Physical Society conference here this week. This marks the third time IBM has doubled the performance of its quantum computing systems since 2017.
Bob Sutor, an IBM vice president overseeing the research efforts for quantum computing and AI, said this latest result puts it on a more predictable path to achieving Quantum Advantage, which could happen in as little as three years.
Taking a break from the conference, Sutor, along with colleague Sarah Sheldon, who leads the quantum performance and metric team, discussed the technical progress IBM has made with its quantum systems the past few years and what the roadmap looks like for delivering commercially available services. He also offered advice on how to prepare for the coming age of quantum computers and when it might make sense for users to house and maintain quantum systems on premises.
What is the significance of achieving the latest Quantum Volume goal?
Bob Sutor: It sharpens the idea of when quantum computing will be ready to do something really interesting. We have yet to give a sense [of quantum computing's] momentum [or] the progress we've made year to year. But this confirms we have doubled the value and performance of quantum systems twice since 2017.
What it means for a roadmap is everyone developing quantum systems now knows we need to double Quantum Volume every year. If we can do this successfully, we can achieve Quantum Advantage within the next decade, which is the point where we can do things significantly faster and with greater accuracy than you can with today's classical computers.
With the real-world impact of quantum computing a decade away, why should users spend time thinking about how to prepare for it now?
Sutor: For the past year, we've been saying we would see Quantum Advantage within a decade. But, really, we hope to see it in the next three to five years. This is because it's not going to be just IBM working on this, but anyone else involved in making quantum systems like scientists and researchers creating new algorithms. It will be a broad collection of experts from across all industries.
If you look at who are members of our Q Network, there are companies like Exxon-Mobil, JPMorgan Chase and various governments. You don't want to wake up some day and find that two of your biggest competitors are using [quantum computing] to gain a 30% improvement over some offering or service you have just because you didn't want to bother looking into it.
IBM has mentioned tools that might be able to take existing applications and data residing on host systems and allow them to work on quantum systems. How far away is that?
Sutor: You can't think of quantum as being a one-for-one replacement with classical applications; quantum is more complementary. It isn't as simple as recompiling an existing app. The approach to programming is completely different. Quantum computers are not big data machines, not the kind of machines you pump a ton of data through and then work with all that data algorithmically.
How important will the upcoming Quantum Computation Center be for the overall advancement of quantum computing in general?
Sutor: The key message that came from the Q System One announcement was it's time now to move quantum computing out of the research labs and into the cloud data centers, meaning there is more to be done than just working on the chip. It really involves all the elements from the chip, all the way up through the systems engineering and how you can actually host [quantum computers].
Fortune 1000 users sometimes wonder, 'How do I know that I am just playing with this crazy thing in a back room someplace?' But what we are and will be working toward [in the new lab] is making this a commercially viable, retainable and upgradeable system, which is what I think the significance of the lab will be.
There is ongoing research on developing quantum systems using silicon, which supposedly would significantly reduce costs associated with maintaining a subzero quantum environment and maybe hasten the arrival of quantum systems in businesses. Is IBM exploring these capabilities?
Bob Sutorvice president of IBM
Sarah Sheldon: There are several implementations for quantum computers being worked on throughout the research community. But our focus is on superconducting qubit systems. They are systems we know how to make well and [uses] the infrastructure and the expertise IBM Research has, especially in terms of fabrication and the research we have done the past many years. We do keep an eye on all research throughout the field.
So, you haven't looked into what it would take to deliver a silicon-based system?
Sutor: I can't speak to what we have actually done, but you can assume that we are very technically aware of all the technologies in progress out there.
Can you see the day when quantum computers can be housed and supported in an IT shop? Or, will quantum services forever be available only as cloud services?
Sutor: Difficult to say, because there could be a very rapid evolution of this machine ... Personally, I'd like to access it over the cloud for a while and have IBM and others improve the performance, rather than see users buy something concrete and have to upgrade it themselves. Leave the development and upgrading of the machine to the vendors until we get to a point where we have this achieved Quantum Advantage.
What advice would you give a Fortune 1000 organization about the best way to get ready?
Sutor: I would suggest they go to YouTube and look at some of our videos, one in particular by Talia Gershon that has 3 million page views, where she explains quantum to a 7-year-old, a teenager, someone in their 20s – five different perspectives on the technology.
If they are interested after that, there are things like the IBM Q Experience, where we give them Qiskit, a free open source development kit where they can start to get involved. From there, you can progress to more formal training classes, as well as sitting down with a consultant who will go over what you [have] done and make recommendations to fill out [what] your roadmap should be.