I think I'll go and buy a monkey instead.
(And a book about algorithms...)

    KreAture

----- Original Message -----
From: "Daniel Serpell" <daniel@ATMLAB.UTFSM.CL>
To: <PICLIST@MITVMA.MIT.EDU>
Sent: Monday, June 23, 2003 5:23 PM
Subject: Re: [PIC]: New version of the PIC Superoptimizer.


> On Mon, Jun 23, 2003 at 07:50:31AM -0700, James Newton, webhost wrote:
> > source= http://www.piclist.com/postbot.asp?id=piclist\2003\06\20\201728a
> >
> > Let me see if I understand:
> >
> > This program contains:
> > a) a small language in which you can define an algorithm
> > b) a program to generate random inputs for that algorithm and record the
> > result
> > c) a PIC simulator
> > d) a program to generate random sequences of PIC instructions
>
> Not random, the sequences are tried exaustively. First, all one inst.
> sequences, then all 2 ins. seq., etc.
>
> > And the program then tries different sequences until it hits upon one
that
> > produces the expected outputs when given the inputs?
>
> Yes.
>
> > So it generates code in the same way that the old joke about the room
full
> > of monkeys randomly typing will eventually produce the works of
Shakespeare?
> >
> > And this actually works? I must not be understanding something.
>
> Yes, it works for small code fragments. You can search thru all 6
> instruction sequences in about an hour in modern computers, 7 inst.
> sequences in a couple of days, 8 inst. sequences in a couple of
> months, etc. If fewer instructions are tried (it's configurable),
> you can go up thru 9 instruction sequences in a reasonable amount
> of time.
>
> Test the examples, and give coments!
>
>     Daniel.
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways.  See http://www.piclist.com/#archives for details.
>

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.