A computer architecture based on quantum mechanics, the science of atomic structure and function. Quantum computing is radically different from ordinary computers ("classical computing"). It can only solve certain problems, all of which are mathematically based and represented as equations. Quantum computer processing emulates nature at the atomic level and one of its more auspicious uses is the analysis of molecular interactions to uncover nature's mysteries. See quantum mechanics
In 1998, the first quantum computers were demonstrated at Oxford University and IBM's Almaden Research Center. By 2020, there were approximately a hundred working quantum computers worldwide. Because of the high cost of building and maintaining quantum computers, quantum computing is likely to be offered mostly as a cloud service rather than machinery for sale to companies. Time will tell. See quantum coprocessor
and quantum cloud
Computations Can Be Staggering
There are many problems that bog down even the fastest supercomputers because numbers tend to grow very fast. Take for example the classic traveling salesman problem, which attempts to find the most efficient round trip between cities. The first step is to compute all possible routes, and if the trip involved 50 cities, the result is a number 63 digits long. Whereas classical computers may take days or even months to solve such problems, quantum computers are expected to have answers in mere minutes or seconds. See quantum supremacy
, quantum teleportation
, binary values
and rice and chessboard legend
Qubit Superposition and Entanglement
Quantum computing uses the "qubit," or quantum bit, comprising one or more electrons, and there are various approaches to their design. Quantum superposition is the condition that allows a qubit to be in multiple states at the same time (see qubit
). Entanglement is the property that allows one particle to relate to another over distance.
Gate model and quantum annealing are the two broad categories of quantum computer architectures.
Gate Model QC
Gate model quantum computers use gates similar in concept to classical computers but with vastly different logic and architecture. Several companies are developing gate model machines, including Google, IBM, Intel, Rigetti and Honeywell, each with different qubit designs. The quantum chip is programmed by sending microwave pulses to the qubits. Digital-to-analog and analog-to-digital conversion takes place at the QC chip.
IBM's Q Experience in the Cloud
In 2016, IBM made a 5-qubit gate model quantum computer available in the cloud to allow scientists to experiment with gate model programming. A year later, the open source Qiskit development kit and a second machine with 16 qubits were added. The IBM Q Experience includes a library of educational materials.
IBM Q - Gate Model
Intel 49-Qubit Quantum Computer
Superconducting materials are used that must be kept at subzero temperatures, and this photo shows the covers removed to expose the quantum chip at the bottom. See superconductor
. (Image courtesy of IBM Research, www.research.ibm.com)
In 2018, Intel announced its Tangle Lake gate model quantum chip with a unique architecture of single-electron transistors coupled together. Intel CEO Brian Krzanich shows the chip at CES 2018. (Image courtesy of Intel Corporation.)
D-Wave Systems in Canada offers the only "quantum annealing" computer. D-Wave computers are huge, refrigerated machines with up to 2,000 qubits that are used for optimization problems such as scheduling, financial analysis and medical research. Annealing is used to find the optimum route or the most efficient combination of settings to solve a problem.
D-Wave Chips Are Cool Too
D-Wave's latest quantum annealing chip has 5,000 qubits. Like gate model quantum computers, a refrigeration system is necessary. Using liquid nitrogen and liquid helium stages from top to bottom, it keeps getting colder all the way down to minus 459 degrees Fahrenheit. (Images courtesy of D-Wave Systems, Inc., www.dwavesys.com)
Algorithms and Error Correction
The algorithms for solving real-world problems must be invented first, because the algorithms exploit the magic (superposition and entanglement) of quantum computers.
The difficulty in designing quantum computers has to do with errors. The qubit interactions cause errors, and error correction is extremely difficult but also key to practical quantum computers. There are hurdles to overcome with both gate model and annealing methods. However, scientists believe everyday quantum computing is inevitable. The general goal is that by 2030, mostly error-free gate-level quantum computers with more than 100 qubits will be available.
A Potential Catastrophe
Eventually, quantum computers are expected to factor huge numbers and should be able to crack cryptographic keys in a matter of seconds. Scientists contend it is only a matter of time before this becomes a reality. When it does, it has menacing implications as every encrypted transaction as well as every existing cryptocurrency system in the world will be vulnerable to hacking. However, quantum-safe methods are also being developed. See quantum secure
Around the World
As of 2022, the countries that are researching and investing in quantum computing are the U.S., Canada, Germany, France, U.K., Netherlands, Russia, China, South Korea and Japan.
Are We at a Similar Stage?
Quantum computing is in the very early stages of development. When an eight-ton UNIVAC I in the 1950s evolved into a chip decades later, it makes one wonder what quantum computers might look like 50 years from now. See UNIVAC I