http://www.perlmonks.org?node_id=425630


in reply to Re^5: RFC: Implicit Parallelization Pragma
in thread RFC: Implicit Parallelization Pragma

I agree. Fibers are an interesting, and potentially useful architectural feature of Win32, but they do not address the OP topic directly.

There are two types of parellelism that need to be addressed.

Just a notion :)


Examine what is said, not who speaks.
Silence betokens consent.
Love the truth but pardon error.
  • Comment on Re^6: RFC: Implicit Parallelization Pragma

Replies are listed 'Best First'.
Re^7: RFC: Implicit Parallelization Pragma
by Anonymous Monk on Jan 27, 2005 at 19:03 UTC

    Indeed, your vision of the sync-point parallelism is probably what the ideas in the original node need as a viable CPU platform.

    I'd love to see the assembly code for that sort of piece-wise programming. This is where Smalltalkish uber-OOP could be actually used on hardware level.

      I'd love to explore these ideas further.

      However, apparently in-depth discussion is beyond the remit of this site. However, if you would like to take this discussion further at another place, you could grab yourself an id for the purpose of our making private contact and we can arrange for that discussion to take place elsewhere.


      Examine what is said, not who speaks.
      Silence betokens consent.
      Love the truth but pardon error.
      Indeed, your vision of the sync-point parallelism is probably what the ideas in the original node need as a viable CPU platform.

      Do Anonymonk's check for follow-ups?

      With regard to viable CPU platforms--did you see Monday's announcement?

      A little more (speculative) information.It comes to pass :)

      Note the diagram "Distributed processing with cells" and the mention of "software cells".

      Boy, would I like to get my mitts on one of those, along a with a properly threaded OS.


      Examine what is said, not who speaks.
      Silence betokens consent.
      Love the truth but pardon error.