Almost all of today’s computers are based on simple Turing Theory and employ Boolean logic based on binary mathematics. Even “parallel” computers are really complex Turing engines employing multiple computing modules which deal with pieces of incoming data (chunks, bytes, instructions, etc.). There has been some research into biological computing using enzymes or large-molecule systems as memory, shift registers, etc., but this has not proven to be very practical.
Quantum computing is based on a different physics than digital computing. Instead of having two (or three) states-per-element like digital computers which are off, on, or neither, quantum computers can have all three states at the same time. An 8-bit digital computer can exist in only one of 256 states at a time while an eight bit quantum computer can exist in all 256 states at a time and theoretically, work on 256 calculations at once (quantum parallelism). Each of the 256 numbers in this 8-bit example has an equal probability of being measured so that a quantum processor functions as a random number generator. The actual register is representing all of these values at once but a single value output only occurs at measurement. While a classical digital computer would have to operate on each number from 0 to 255, quantum computers require only one pass through the “processor”, radically reducing calculation time. Of course, the larger the register size, the larger the number – even a simple 10-bit quantum computer could scream past a supercomputer.
Where the digital computer uses binary digits (bits), the quantum computer uses qubits, but qubits are extremely difficult to generate. A quantum switch must be undisturbed. Light, molecules, or impinging fields required for the proper operation of a quantum computer depend on the interaction of the various qubits without any outside influence. When disturbed, the qubit becomes quite Newtonian rather than quantum and selects a definite state – by chance becoming, dare we say, digital.
Paul Benioff of the Argonne National Laboratory first applied quantum theory to computers in 1981 and David Deutsch of Oxford proposed quantum parallel computers in 1985, years before the realization of qubits in 1995. Qubits are made using various techniques. A group at the national Institute of Standards in Boulder trapped a single atom with missing electrons (an ion) with two energy levels by containing it with magnetic and electric fields at -273 degrees C. Another group at California Institute of Technology made qubits from polarized light using a device which allows photons to interact while they pass though a stream of cesium atoms interacting in an XOR-like manner. At Los Alamos, researchers make qubits by trapping ions. Ion traps housing up to six ions have already been produced – far short of the thousands required for a useable quantum computer.
Although quantum computers have not been built to date, many of the mechanisms required like error correction and algorithm construction are being investigated. Because of their multiple states (unlike two-state digital processes) quantum computers will have some of the problems that analog computers had – namely error correction and calculation reliability (although this does not sound very quantum mechanical). Physicists are arguing what type of error correction will work with qubits and quantum measurement in general. Believe it or not, John von Neuman’s work in computer error detection and correction is being re-examined and has led to new efforts in quantum error correction.
DARPA, funded a $5 million dollar Quantum Information and Computing Institute for the purpose of investigating quantum computing and its applications. It may be well into the 21st century before we see quantum computing used at a commercial level but there is little doubt that the research on new forms of calculation and error correction will improve the state of digital computing, data compression, and error correction.
Elmer Smalling III
Copyright ©1999 Elmer Smalling III