5 Questions to Decide Which Quantum Computer
The race to define the potential value of quantum computers is heating up. More and more organizations are creating quantum innovation groups to explore exciting opportunities for quantum systems. The competitive advantage quantum computing offers is compelling.
Given the dramatic differences between quantum and classical hardware and software development, it’s critical to start evaluating your path to quantum, right now.
The challenge is that a quantum computer isn’t scalable enough today to solve large production problems. So, we all must learn how to explore these systems in new and innovative ways. We can’t just load up the classical application and check it out. We have to think and work differently.
Things to Know About That Quantum Computer
Quantum is as different from classical computing as an abacus is to an iPhone. Seriously. Because of that delta, there is so much to learn. Here are a few tidbits you need to understand as you begin your journey on the path to quantum.
Every quantum computer is based on slightly different hardware technology, using diverse types of qubits. Testing a single quantum computer will not give you the knowledge or insights you need to truly explore the quantum potential.
This diversity means that different systems will be better at solving different types of problems. So, you need to be ready to think about leveraging a different quantum computer for different needs.
Quantum software has to be coded to run on a specific computer. This means your diverse quantum computers will demand diverse software requirements. Since quantum experts are the only ones who can use the software development kits (SDKs) to write quantum software, that’s a significant investment of time and money just to explore quantum computing. Which is why you want to look at ready-to-run, hardware-agnostic software and not just invest massive dollars to start writing programs.
While a quantum computer can’t solve that production problem today, there are valid mathematical approaches to measuring and testing smaller problem models to begin to understand the value you can expect. BTW, this is not the same as running a classical algorithm on a QPU simulator.
Questions About Your Quantum Infrastructure
As you begin to plan for exploring quantum, there are some questions you need to ask yourself to plan for the best possible strategy. Here are a few questions to consider.
Which quantum computers are the best match for your problem?
As I mentioned, different challenges will likely require a different type of quantum computer. Especially today. The diverse qubit architectures offer different approaches to solving computations and problems, with varying degrees of potential performance, results and accuracy. That means only testing one QPU leaves a lot to chance.
You’ll need to match the priority problems you defined as quantum-possible so you can be sure you explore your best opportunities. This means you need to review the problem itself regarding its requirements for processing and mathematical computations, then map those requirements to the capabilities of a variety of QPUs to be sure you test the best quantum computers for your specific needs.
How are you going to measure and test these systems?
As with any benchmarking or testing project, you’ll need a plan regarding what specifics you want to test. This comes from the problems themselves, meeting key requirements, the current benchmarks you’d like to compare and other key factors. Remember, performance isn’t the focus yet, accuracy and depth of results are factors you can focus on for today.
Also, be ready to expand your expectations. Quantum computers deliver different results than classical systems. For example, with optimization problems, a quantum computer delivers a diversity of results that all meet the optimization requirements. Classical systems only deliver the top result. Quantum systems give you the opportunity for a deeper set of insights since you can examine a variety of results that may have the tiniest deviations, yet one may match your immediate situation better than the optimum single answer from classical.
How are you going to select the software/algorithms you need?
It would be easier if you could simply run a classical algorithm on a quantum computer and measure it. But that’s not going to be very helpful. You see, you need to write new programs and algorithms to run a quantum computer.
The challenge is that today’s SDKs are complex. It can take a year to develop an entire program for a single quantum computer type. Which is why many are taking the approach of testing a single algorithm to simplify their exploration. We don’t think that’s the best option.
Ready-to-run software gives you the opportunity to run your problem on the quantum computers you’ve selected, without the level of investment needed to hire quantum experts to create programs so you can fully explore quantum.
Take some time to define your best software option rather than just jumping to follow what the general market is doing in these early stages. Your business will benefit.
How much money, time and risk are you willing to invest in your first exploration?
Yes, quantum computing time is expensive. As is software development, if you choose to create complex programs for your business at this point in time.
As always, there are ways to limit the risk and investment you need to make, even as you commit to a comprehensive and strategic exploration of your path to quantum.
Think strategically, work with experts who are vendor-neutral, ask yourself WHY you need to do a specific thing now, rather than later.
For example:
Why would you spend an inordinate amount of time developing a software program when you are in your evaluation period?
Why would you hire a team of quantum experts to run an algorithm on a quantum computer?
What is the best approach to explore your quantum-potential without making significant investments, with significant meaning millions and millions of dollars?
The Bottom Line
In every advanced technology market, the early days are filled with excitement and challenges. We naturally tend to follow the vendors who are leading the market. After all, they have the most knowledge.
With quantum, that may or may not be the best path. Everyone is learning, and no one is the complete expert. Following the rest of the market may not be the best strategy in this case.
Yes, there are vendor-neutral experts who can help you. But you also need to ask questions yourself, questions that bring the exploration of quantum directly into alignment with your business requirements. A quantum computer is meant to drive your business, not the market’s science experiments. Right?