4 ms·

Do we have contenders to challenge integrated circuits? 3D transistor[0] or graphine transistors[1]? I just hope the next revolutionary step happens in my life

Do we have contenders to challenge integrated circuits? 3D transistor[0] or graphine transistors[1]? I just hope the next revolutionary step happens in my lifetime. Not complaining about all the stuff already happening, just want more! Edit: [3] ------------------------ [0] https://en.wikipedia.org/wiki/Multigate_device https://en.wikipedia.org/wiki/Multigate_device [1] https://phys.org/news/2016-05-graphene-based-transistor-clock-processors.html https://phys.org/news/2016-05-graphene-based-transistor-cloc... [3] https://en.wikipedia.org/wiki/Tunnel_field-effect_transistor https://en.wikipedia.org/wiki/Tunnel_field-effect_transistor

deepnotderp 10y ago
The problem is that everything useful has either showed or stopped scaling. For example, at 10nm (7nm for TSMC) we're getting an ungodly number of logic gates, but sram cell size scaling and Dennard scaling, two things we actually care about have slowed. Similarly, the cost of data movement hasn't come down much. So 3d transistors and graphene transistors won't help all too much. And besides, FinFET s are basically a form of 3d transistors.
- smaddox 10y ago
  Logic cell density is still increasing approximately exponentially [1]. [1] https://newsroom.intel.com/newsroom/wp-content/uploads/sites/11/2017/03/Kaizad-Mistry-2017-Manufacturing.pdf https://newsroom.intel.com/newsroom/wp-content/uploads/sites...
  - deepnotderp 10y ago
    Exactly! And that's what we don't care about.
    - smaddox 10y ago
      Okay, fine, logic and SRAM cell density are still shrinking (at Intel). Now are you happy? Just read the report.
acjohnson55 10y ago
I took a class a long time ago that was about contenders for computing on different logical substrates, which might be enabled by active areas of research in the somewhat-distant future: https://registrar.princeton.edu/course-offerings/course_details.xml?courseid=010142&term=1112 https://registrar.princeton.edu/course-offerings/course_deta...

deepnotderp 10y ago

The problem is that everything useful has either showed or stopped scaling. For example, at 10nm (7nm for TSMC) we're getting an ungodly number of logic gates, but sram cell size scaling and Dennard scaling, two things we actually care about have slowed. Similarly, the cost of data movement hasn't come down much. So 3d transistors and graphene transistors won't help all too much. And besides, FinFET s are basically a form of 3d transistors.

smaddox 10y ago
Logic cell density is still increasing approximately exponentially [1]. [1] https://newsroom.intel.com/newsroom/wp-content/uploads/sites/11/2017/03/Kaizad-Mistry-2017-Manufacturing.pdf https://newsroom.intel.com/newsroom/wp-content/uploads/sites...
- deepnotderp 10y ago
  Exactly! And that's what we don't care about.
  - smaddox 10y ago
    Okay, fine, logic and SRAM cell density are still shrinking (at Intel). Now are you happy? Just read the report.

acjohnson55 10y ago

I took a class a long time ago that was about contenders for computing on different logical substrates, which might be enabled by active areas of research in the somewhat-distant future: https://registrar.princeton.edu/course-offerings/course_details.xml?courseid=010142&term=1112 https://registrar.princeton.edu/course-offerings/course_deta...