Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.python > #8219

Re: How can I speed up a script that iterates over a large range (600 billion)?

References <f97b3159-318a-41e2-9f28-339fa4a81d79@k6g2000yqc.googlegroups.com> <its0ar02sia@news2.newsguy.com> <4e01d983$0$29977$c3e8da3$5496439d@news.astraweb.com>
From Ian Kelly <ian.g.kelly@gmail.com>
Date 2011-06-22 09:52 -0600
Subject Re: How can I speed up a script that iterates over a large range (600 billion)?
Newsgroups comp.lang.python
Message-ID <mailman.285.1308757993.1164.python-list@python.org> (permalink)

Show all headers | View raw


On Wed, Jun 22, 2011 at 6:01 AM, Anny Mous <b1540457@tyldd.com> wrote:
> def sieve():
>    """Yield prime integers efficiently.
>
>    This uses the Sieve of Eratosthenes, modified to generate the primes
>    lazily rather than the traditional version which operates on a fixed
>    size array of integers.
>    """
>    # This implementation based on a version by Melissa E. O'Neill,
>    # as reported by Gerald Britton:
>    # http://mail.python.org/pipermail/python-list/2009-January/1188529.html
>    innersieve = sieve()
>    prevsq = 1
>    table  = {}
>    i = 2
>    while True:
>        if i in table:
>            prime = table[i]
>            del table[i]
>            nxt = i + prime
>            while nxt in table:
>                nxt += prime
>            table[nxt] = prime
>        else:
>            yield i
>            if i > prevsq:
>                j = next(innersieve)
>                prevsq = j**2
>                table[prevsq] = j
>        i += 1

This appears to have a small bug in it, but perhaps it doesn't matter.
 At the "yield i" statement, it is possible that i > prevsq.  It could
be that i == next(innersieve)**2, in which case i is yielded as prime
despite being composite.  This would then cause additional errors
further along.

The only way this could happen would be if there were two consecutive
primes such that all the numbers between their squares are composite,
thereby failing to add the next prime to the table until after its
square has been reached.  This seems a rather unlikely scenario, but I
can't say for certain that it never happens.

The Haskell version does not contain this flaw, as far as I can tell.

Cheers,
Ian

Back to comp.lang.python | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

How can I speed up a script that iterates over a large range (600 billion)? John Salerno <johnjsal@gmail.com> - 2011-06-21 12:48 -0700
  Re: How can I speed up a script that iterates over a large range (600 billion)? Ian Kelly <ian.g.kelly@gmail.com> - 2011-06-21 14:02 -0600
  Re: How can I speed up a script that iterates over a large range (600 billion)? Irmen de Jong <irmen@-NOSPAM-xs4all.nl> - 2011-06-21 22:10 +0200
    sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Irmen de Jong <irmen@-NOSPAM-xs4all.nl> - 2011-06-21 22:22 +0200
      Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? John Salerno <johnjsal@gmail.com> - 2011-06-21 14:09 -0700
        Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Irmen de Jong <irmen.NOSPAM@xs4all.nl> - 2011-06-21 23:39 +0200
        Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Ian Kelly <ian.g.kelly@gmail.com> - 2011-06-21 15:41 -0600
          Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? John Salerno <johnjsal@gmail.com> - 2011-06-21 14:48 -0700
        Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Vlastimil Brom <vlastimil.brom@gmail.com> - 2011-06-21 23:33 +0200
        Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Paul Rubin <no.email@nospam.invalid> - 2011-06-21 17:05 -0700
          Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? John Salerno <johnjsal@gmail.com> - 2011-06-21 18:21 -0700
            Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Paul Rubin <no.email@nospam.invalid> - 2011-06-21 19:09 -0700
              Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? John Salerno <johnjsal@gmail.com> - 2011-06-21 20:02 -0700
                Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Irmen de Jong <irmen.NOSPAM@xs4all.nl> - 2011-06-22 19:46 +0200
            Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? MRAB <python@mrabarnett.plus.com> - 2011-06-22 03:10 +0100
            Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? Mel <mwilson@the-wire.com> - 2011-06-21 23:02 -0400
              Re: sorry, possibly too much info. was: Re: How can I speed up a script that iterates over a large range (600 billion)? John Salerno <johnjsal@gmail.com> - 2011-06-21 20:41 -0700
  Re: How can I speed up a script that iterates over a large range (600 billion)? Mel <mwilson@the-wire.com> - 2011-06-21 16:19 -0400
    Re: How can I speed up a script that iterates over a large range (600 billion)? Tim Roberts <timr@probo.com> - 2011-06-22 23:21 -0700
  Re: How can I speed up a script that iterates over a large range (600 billion)? MRAB <python@mrabarnett.plus.com> - 2011-06-21 21:26 +0100
  Re: How can I speed up a script that iterates over a large range (600 billion)? Terry Reedy <tjreedy@udel.edu> - 2011-06-21 19:30 -0400
    Re: How can I speed up a script that iterates over a large range (600 billion)? Paul Rubin <no.email@nospam.invalid> - 2011-06-21 17:00 -0700
      Re: How can I speed up a script that iterates over a large range (600 billion)? Terry Reedy <tjreedy@udel.edu> - 2011-06-22 00:18 -0400
        Re: How can I speed up a script that iterates over a large range (600 billion)? Paul Rubin <no.email@nospam.invalid> - 2011-06-21 22:32 -0700
          Re: How can I speed up a script that iterates over a large range (600 billion)? Terry Reedy <tjreedy@udel.edu> - 2011-06-22 18:46 -0400
  Re: How can I speed up a script that iterates over a large range (600 billion)? Benjamin Kaplan <benjamin.kaplan@case.edu> - 2011-06-21 20:07 -0700
  Re: How can I speed up a script that iterates over a large range (600 billion)? Chris Torek <nospam@torek.net> - 2011-06-22 05:58 +0000
    Re: How can I speed up a script that iterates over a large range (600 billion)? Paul Rubin <no.email@nospam.invalid> - 2011-06-21 23:23 -0700
      Re: How can I speed up a script that iterates over a large range (600 billion)? Slaunger <slaunger@gmail.com> - 2011-06-23 02:52 -0700
    Re: How can I speed up a script that iterates over a large range (600 billion)? Ian Kelly <ian.g.kelly@gmail.com> - 2011-06-22 01:16 -0600
    Re: How can I speed up a script that iterates over a large range (600 billion)? Anny Mous <b1540457@tyldd.com> - 2011-06-22 22:01 +1000
      Re: How can I speed up a script that iterates over a large range (600 billion)? Chris Angelico <rosuav@gmail.com> - 2011-06-22 22:28 +1000
      Re: How can I speed up a script that iterates over a large range (600 billion)? Ian Kelly <ian.g.kelly@gmail.com> - 2011-06-22 09:52 -0600
    Re: How can I speed up a script that iterates over a large range (600 billion)? MRAB <python@mrabarnett.plus.com> - 2011-06-22 16:33 +0100

csiph-web