Is there a limit to Moore's law?

Replies

• 

  Kaustubh Katdare

  #-Link-Snipped-# - Thanks a lot for the topic. I had the same question few years ago when I was first introduced to the Moore's law. Intel's transistors are going 22nm and have the power to turn on/off a 100 billion times in a second! It's however still in sync with the Moore's law. Here's an interesting page on Intel's website: Moore’s Law: Fun Facts about the topic.
  
  Talking about the limit, I found an interesting paper (article) on IEEE website that says year 2036 is going to be the ultimate year for the Moore's law when we'll max out the number of transistors we can accommodate on a chip - and it's defined by Compton Wavelength. If electrons were implemented as the smallest units to act as transistors, the Moore's law maxes out in year 2036. Refer to the source below for computation and other details.
  
  Source: #-Link-Snipped-#
  
  I'd however not place any bets on the law. The pace at which electronics is advancing is mind boggling. I still don't know what to do with all the processing power my phone offers to me.
• 

  Ankita Katdare

  Here's a discussion on CrazyEngineers about a related topic - Electronics Development After Moore's Law ?
• 

  Ramani Aswath

  Quantum Computing may put the seal on this law:
  #-Link-Snipped-#
  
  It is like the chain letters. Send to 9 people. Within 11 cycles the population of earth will be exhausted if no repeats are allowed.
  Likewise I guess once we reach to atomic dimensions, Moore's Law will have to saturate theoretically.
• 

  Smiech

  Thanks Kaustubh, Ankita, A.V. Ramani for your Replies and valuable information.
  Any way, the ultimate limit of the MOS gate-drain length depends not on the technology but rather on physics itself.
  What does appear certain is that, for several decades at least, there will not emerge a technique that can compete with silicon MOS technology.
  There is no doubt that, with new chip-manufacturing technologies, together with improved architecture systems, the silicon-based MOS will remain the cornestone of the semiconductor industry.
  Another possibility is to build transistors using nanotubes, which are very good conductors. Carbon is not silicon. Right now, nanotubes are difficult to handle, but I guess, these problems will not last for long.
  Maybe, the most promising alternative is the integration of silicon microelectronics with photonic microelectronics. Such a system would waste less energy than its chip predecessor and would have easier connections, making it possible to build large-scale integrated circuits. Plus, the higher bandwidth of optical interconnections greatly increases the speed of the microprocessor.
  What do you think about this?
• 

  Ramani Aswath

  The biological brain is a gel like substance that packs remarkable performance /cc. Works with an electro-chemical sort of info transfer and processing. Very power efficient. Housed in a strong housing and floating in a fluid it is quite shock resistant.
  Maybe a biomimetic gel like processor will emerge some day.