Page 2 of 2

Posted: Sat Apr 16, 2011 2:51 pm
by Arlet
I've read that the 6502 feature size was 6 um, and the die size 21 sq mm. On a modern process with say 45 nm feature size, you could fit about 1000 CPUs on a sq mm.

With a 100 sq mm chip, that would be about 100,000 CPUs.

Of course, without integrated memory and I/O buses, this would be a completely useless exercise.

Posted: Sat Apr 16, 2011 3:13 pm
by e2020
are you serious? are these numbers correct?

Posted: Sat Apr 16, 2011 4:02 pm
by Arlet
Another way to look at it... the 6502 had about 3500 transistors, while the biggest moderns chips have a few billion. Almost a million times as much.

Posted: Sat Apr 16, 2011 4:59 pm
by ChuckT
e2020 wrote:
are you serious? are these numbers correct?
All you have to do is open up an old Apple II or Commodore 64 and compare the chip against a microprocessor like Atmel, Microchip or Parallax. The newer chips are more powerful and much smaller because they use a smaller nanometer process.

Posted: Sat Apr 16, 2011 6:11 pm
by ChuckT
e2020 wrote:
thanks for asking there!
http://www.technologyreview.com/computi ... ?p1=A1&a=f

In the future, FPGAs will be able to make hardware changes and not just software changes so you may be able to go to a higher speed and have a different clock for the bus and the screen.