COMPUTING COMPUTERS TAKING A QUANTUM LEAP Quantum computers will harness the power of atoms and molecules to perform calculations billions of times faster than today’s silicon-based computers. They will also enable new revolutionary applications � ASHUTOSH BHATIA tween your classical computer and a But can it continue forever? quantum computer. Whereas the clas- The basic processing unit in a com- T hink of a computer whose sical computer obeys the well-under- puter chip is the transistor, which acts memory is exponentially larger stood laws of classical physics, the like a small switch. The binary digits than its apparent physical size. quantum computer uses physical phe- ‘0’ and ‘1’ are represented by the tran- Or a computer that can manipulate an nomenon unique to quantum mechan- sistor being turned off or on, exponential set of inputs simulta- ics (especially quantum interference) respectively. Currently, thousands of neously. The quantum computer to realise a fundamentally new mode electrons are used to drive each tran- would be one such computer. Rela- of information processing. sistor. As the processing power in- tively few and simple concepts from creases, the size of each transistor re- Moore’s law and the future quantum mechanics are needed to duces. of computers make quantum computers a possibil- If Moore’s law continues unabated, ity. In 1965, Intel’s cofounder Gordan each transistor would be as small as a Quantum computers open up a Moore noted that the processing hydrogen atom by the year 2030. At new era for high-speed computations. power (number of transistors and this size, the quantum nature of elec- They will be 1,000,000,000 times faster speed) of computer chips was dou- trons in the atoms becomes significant than current silicon-based computers. bling every 18 months or so. This trend and generates errors in the computa- Today’s high-speed computer sitting has continued for nearly four decades. tion. in front of you is fundamentally no However, it is possible to exploit different from its 30-tonne ancestors, the quantum physics as a new way to which were equipped with some do computation. And this new way 18,000 vacuum tubes and 805 opens up fantastic new computational kilometres (500 miles) of wiring! power based on the wave nature of Although computers have become quantum particles. Figs 1 and 2 show more compact and considerably faster the size and number of transistors over in performing their tasks, the task re- time scale up to 2030, respectively. mains the same: manipulate and in- Particle-wave duality terpret an encoding of binary bits into a useful computational result. A bit is We normally think of electrons, atoms Fig. 1: Size of transistors in a computer chip the fundamental unit of information, and molecules as particles. But each by the year 2030 classically represented as a ‘0’ or ‘1’ in of these objects can also behave as your digital computer. waves. This dual particle-wave Each classical bit is physically behaviour was first suggested in the realised through a macroscopic physi- 1920s by Louis de Broglie. cal system, such as the magnetisation This concept emerged as on a hard disk or the charge on a ca- follows: Thomas Young’s experiment pacitor. A document, for example, with double slits in the early 1800s comprising ‘n’ characters stored on the shows that light behaves as a wave. hard drive of a typical computer is ac- But Einstein’s explanation of the pho- cordingly described by a string of 8n toelectric effect in the year 1905 shows 0’s and 1’s. that light consists of particles. In 1923, Fig. 2: Number of transistors in a computer Herein lies a key difference be- de Broglie suggested this dual particle- chip by the year 2030 44 • MAY 2005 • ELECTRONICS FOR YOU W W W . E F Y M A G . C O M
COMPUTING wave property might apply to all par- bility opens up for at- ticles including electrons. Then in 1926, oms. Electrons have a Davisson and Germer found that elec- wave property that al- trons scattered off a nickel crystal be- lows a single electron to haved as waves. Since then neutrons, be in two orbits simul- atoms and even molecules have been taneously. In other shown to behave as waves. The waves words, the electron can Fig. 5: When a single photon encounters a beam splitter, the tell us where the particle is likely to be in a superposition of photon emerges in a superposition of the reflected path and the be found. both orbits. transmitted path Fig. 4 shows two at- Bits and qubits oms, each with a single The basic data unit in a conventional electron in a superposi- (or classical) computer is the bit, or tion of two orbits. Each binary digit. A bit stores a numerical atom represents binary value of either ‘0’ or ‘1.’ An example digits ‘0’ and ‘1’ simul- of how bits are stored is given by a taneously. The two at- CD-ROM, where ‘pits’ and ‘lands’ (ab- oms together represent sence of a pit) are used to store the the binary numbers ‘00,’ binary data. We could also represent a ‘01,’ ‘10’ and ‘11’ simul- bit using two different electron orbits taneously. in a single atom. In most atoms, there To distinguish this are many electrons in many orbits. But new kind of data stor- we need to consider only the orbits age from a conventional available to a single outermost elec- bit, it is called a quan- tron in each atom. tum bit, or qubit. Each Fig. 6: (a) Two super-imposed water waves ‘A’ and ‘B’ and (b) Fig. 3 shows two atoms represent- atom in Fig. 4 is a qubit. the sum of ‘A‘ and ‘B’ waves looks like neither of its components ing the binary number ‘10.’ The inner The key point is that a orbits represent digit ‘0’ and the outer qubit can be in a superposition of the the same time, but as waves, they can. orbits represent digit ‘1.’ The position digits ‘0’ and ‘1.’ Superposition states Superposition in waves. Fig. 6(a) of the electron gives the number stored allow many computations to be per- shows two superimposed waves ‘A’ by the atom. formed simultaneously, and give rise and ‘B.’ If we were to add these waves However, a completely new possi- to what is known as quantum paral- together numerically, the result lelism. (S=A+B) would be a wave that looks Another example of a qubit is a like neither of its components (see Fig. photon (a particle of light) travelling 6(b)). However, one could retrieve ei- along two possible paths. Consider ther wave by subtracting out the other. what happens when a photon encoun- (The waveform ‘S’ is shown as dotted ters a beam splitter. A beam splitter is to indicate that it is only the apparent just like an ordinary mirror, however waveform; the actual waveform is a the reflective coating is made so thin combination of two different waves. that not all light is reflected and some In the quantum world, the combined light is transmitted through the mir- waveform is a set of amplitude prob- Fig. 3: Two atoms representing the binary ror as well. When a single photon en- abilities.) number ‘10’ counters a beam splitter, the photon Superposition in link list pointers. emerges in a superposition of the re- For an example germane to computer flected path and the transmitted path. programming, one may look at a data One path is taken to be the binary digit structure called the linked list. Each ‘0,’ and the other path is taken to be date node in the list contains a refer- the digit ‘1.’ The photon is in a super- ence, or pointer, to the next data node. position of both paths and so repre- The program traverses the list by sents both ‘0’ and ‘1’ simultaneously. jumping to the next data node indi- cated by the pointer. In a doubly Superpositioning linked list, the data node contains two Fig. 4: Two atoms, each with a single electron Superpositioning means that two pointers, one for traversing to the top in a superposition of two orbits. Each atom things overlap without interfering with of the list and the other for traversing represents binary digits ‘0’ and ‘1’ simultaneously. The two atoms together each other. In classical computers, elec- to the bottom of the list. represent the binary numbers ‘00,’ ‘01,’ ‘10’ trons cannot occupy the same space at Another way of implementing a and ‘11’ simultaneously 46 • MAY 2005 • ELECTRONICS FOR YOU W W W . E F Y M A G . C O M
Recommend
More recommend
