When you look back at the history of computers and how they evolved, 2019 will be regarded as a special year. For that¡¯s when quantum computers started to feel real and not just something out of a sci-fi novel.
One of the companies leading the charge on commercializing quantum computers before anyone else in the world is of course Intel. And the guy tasked with making it happen is James ¡°Jim¡± Clarke, who¡¯s the director of Quantum Hardware research group within Intel¡¯s Components Research Organization.
If you will use a quantum computer on your work desk or in your basement in the future, it would¡¯ve been thanks to the pioneering work Jim Clarke and his team is undertaking at Intel right now, trying to enable the very cutting edge of tomorrow¡¯s technology. Not just bit and byte, but figuring out the scalability of qubit -- which is essentially at the heart of the quantum computing problem.
¡°So in the quantum computing community today, you hear a lot about small quantum systems of between 20 and 50 qubits, and this is certainly very interesting. But to do something significant, to be able to solve a practical application with a quantum computer, we need to have thousands if not millions of qubits,¡± says Jim Clarke. ¡°And we're a long way away from having the number of qubits required to do something that's going to change the world.¡±
This is because right now qubits are controlled in a brute force manner, we have one wire or two wires for every single qubit -- a setup that¡¯s simply impossible to scale into larger, practical system sizes, according to Jim Clarke. He¡¯s less concerned about the very small qubit systems that are cute to play with but impractical to use, and more interested in scaling them into large systems.
Unlike your desktop PC or laptop¡¯s processor (whether it¡¯s Intel or AMD) -- which is matchbox-sized, packed with billions and billions of transistors, that fundamentally work on two bits, zero and one, off and on -- a quantum computing chip¡¯s fundamental unit is a qubit (short for quantum bit), which simply put is a combination of zero and one at any given instant. If you find this confusing, wait till you know the state in which quantum computers exist right now -- suspended inside cryogenic refrigerators at -273 degree celsius, which is when individual qubits can be studied and controlled.
Imagine putting this setup on your lap or your living room -- it¡¯s practically impossible right now. Jim Clarke -- who has co-authored more than 50 papers and has several patents to his name on this subject -- knows this better than anyone else, of course.
¡°That¡¯s why we have begun to fabricate custom design control chips that can sit within the cryogenic refrigerator and control many qubits at the same time in a more scalable manner,¡± Jim Clarke says while referring to Intel Horse Ridge, a first-of-its-kind mixed-signal System-on-Chip that Intel announced in Dec 2019, which reduces the complexity of having hundreds of cables and wires needed to control every single qubit into a single, unified package that sits inside the cryogenic refrigerator and extremely close to the quantum device.
Apart from Intel, Google and IBM are working hard on making quantum computers more practical as well -- both companies posted their respective research breakthroughs in 2019, too. However, Intel¡¯s legacy in chip design and manufacturing is expected to give it an edge over all others.
Not only is the characterization of qubits actually very difficult, according to Jim Clarke, but for many of the main technologies and validation, these qubits have to be operated at very, very cold temperatures. ¡°It¡¯s like we take our quantum chip, or Google or IBM takes their quantum chip, and they put in a dilution refrigerator and study it at let's say 10 millikelvin temperature, which is much colder than the freezing cold experienced even in outer space,¡± says Jim Clarke.
As things stand right now, getting results on the qubit is an extremely slow process, Jim tells me, not to mention the super cold operating environment of the qubit itself that makes scalability extremely challenging. ¡°So what we've done at Intel is we're working with two companies based in Finland to develop basically what we¡¯re calling a cryoprobe. It operates at 1 Kelvin or 2 Kelvin. It's a combination of a dilution refrigerator and a probe station. And we think that we're going to be able to get our information at low temperature roughly 100 to 1000 times faster than what we do today,¡± according to Jim Clarke.
¡°Another big differentiator between Google, IBM and Intel is the type of qubit we are studying. Our qubit looks a lot like a conventional transistor, and conventional transistors are extremely smaller than a superconducting qubit (used by Google and IBM) -- our spin qubit technology is roughly a million times smaller than a superconducting qubit,¡± claims Jim Clarke, ¡°So the hope is that we could scale this technology a lot better than the competition.¡±
Of course, it doesn¡¯t take a genius to figure out that Intel has billions dedicated for semiconductor processing. ¡°So if we combine Intel's capabilities in this space, along with qubit technology that looks a lot like transistors, we think we're going to we think we're going to win in the quantum computing space,¡± feels Jim Clarke.
Despite all the healthy optimism, Jim Clarke¡¯s probably the first to throw caution to the wind, underscoring the importance to recognise that we're still very early in the race to deliver a practical quantum computer. And he¡¯s right.
If you walked into a quantum computing lab, you would see the refrigerator, you would see racks of electronic boxes. It isn¡¯t unlike something you might find in a university electronics lab, Jim tells me. Right now, those racks and racks of electronics are really only able to control a few qubits -- which means this model is not scalable. Similar to the early days of computing based on vacuum tube times, where you had a machine that performed a basic calculation but it took up a whole room¡¯s worth of circuitry to be able to do that simple function.
The first transistor was built in 1947, the first integrated circuit appeared in 1958. The first integrated circuit only had five transistors in it, and it was a very small circuit,¡± Jim Clarke begins to draw a parallel between the evolution of quantum computing against the history of solid state electronics. The first microprocessor came around in 1971, the Intel 4004 microprocessor, and that only had 2500 transistors -- contrast that with today¡¯s Intel Core i9-9900K processor which has roughly over 1.7 billion transistors.
¡°So if I had to plot the journey of quantum computing against the evolution of solid state electronics, we're probably just past the stage of the first integrated circuit. So in equivalent years it might be about 1960,¡± according to Jim Clarke. If you had to go back to the quantum computing lab environment, all that¡¯s Horse Ridge is doing is replacing all the racks of electronics and putting it on a single chip that can go to the bottom of a dilution refrigerator.
There¡¯s no doubt about the promise of quantum computing. Google¡¯s claim in October 2019 of achieving "quantum supremacy" where it solved a complex computation in 200 seconds, which otherwise would have taken the most powerful supercomputers in the world approximately 10,000 years to finish, says everything you need to know about what quantum computers can achieve right now. But they¡¯re still research projects, and there¡¯s a long way before the technology becomes mainstream for kids to use in their basement.
We¡¯re still some time away from a breakthrough moment, as far as quantum computing catching everyone¡¯s imagination is concerned, feels Jim Clarke. ¡°I would say we probably have 10 more years before we have a chip that is going to change our life."
If there¡¯s one thing synonymous with Intel over the past 50 years is Moore¡¯s Law and how it has guided the development of semiconductor-based technology, enabling the digital world we are living in right now. When the dust settles down on quantum computers, will we be mentioning a hypothetical Jim Clarke¡¯s Law in a similar vein?
¡°I don¡¯t think anyone¡¯s asked me about that before, it¡¯s flattering and somewhat scary at the same time,¡± Jim Clarke laughs in response before saying, ¡°If we ever reach that point, I hope that we will be looking back at how Intel has enabled the quantum computing era for all of us. That¡¯s all I hope for.¡±