Search for 'descendants' of Moore's law

When this law is expected to be outdated in 10-15 years, scientists need funds of tens of millions of dollars to find silicon replacement technology and chip production that run faster and work better. than.

The National Science Foundation (NSF) has proposed that the government spend $ 20 million for the 2009 fiscal year and studies will focus on quantum computers, carbon nanotubes, multicore computers . ' progress in the economy and life in the last 20 years depends greatly on the ability to process information, 'said Michael Foster, director of NSF's computing and communications department. 'If Moore's law is no longer true, it means that the silicon chip manufacturing industry cannot grow and must find another way now.'

The 'traditional' way to improve chip performance is to reduce the distance between two transistors because more transistors on an area will perform more on / off commands. However, Moore's law states that 'the number of transistors (transistors) placed in each chip, synonymous with processing capacity, will double in cycles of 18-24 months' will come to a limit of This transistor cannot be smaller anymore and the distance between them cannot be further narrowed.

'Candidates' like nanoscale carbon tubes can produce smaller transistors than current silicon. Foster explains: 'Carbon nanotubes are also used as connections on electronic circuits because they can be sequentially'. However, a complete architecture for this material needs additional investment.

In addition, quantum computers are also a solution to improve processing power. This technology uses atoms and molecules to handle massive amounts of work at great speeds because the basic qubit data unit can carry a value of 0 or 1 simultaneously. In addition, the parallel programming mechanism allows several computers of this type to work together with one program.

Even so, silicon chip makers recognize this and they are thinking of different ways to increase processing power such as placing hundreds of cores on a microprocessor or designing a variety of different arrangements to optimize the system. Therefore, the 'death' of silicon chips can come more slowly.