6!:2 'p: i.1e6'
0.280095
t=: p: i.1e6
(":t) 1!:2 <'\junk\x'
6!:2 '". (1!:1) <''\junk\x'''
1.06673
----- Original Message -----
From: "R.E. Boss" <[email protected]>
Date: Monday, April 20, 2009 11:58
Subject: Re: [Jprogramming] table of primes to 1,000,000,000 or more
To: 'Programming forum' <[email protected]>
> Well, if that is so likely, it must be easy to give an example.
> For which reasonable large interval is it faster to generate
> than to read
> all primes?
>
>
> R.E. Boss
>
>
> > -----Oorspronkelijk bericht-----
> > Van: [email protected] [mailto:programming-
> > [email protected]] Namens Zsbán Ambrus
> > Verzonden: maandag 20 april 2009 19:37
> > Aan: Programming forum
> > Onderwerp: Re: [Jprogramming] table of primes to 1,000,000,000
> or more
> >
> > On Sun, Apr 19, 2009 at 6:53 PM, R.E. Boss
> <[email protected]> wrote:
> > >> Also, a table of primes is never worth to download or
> store, as it's
> > >> faster to generate them online than reading them.
> >
> > > This "never" seems incorrect, assumed reading is linear (in
> the number
> > of
> > > primes) and generating is not.
> > > How long took it to generate the largest known prime?
> Definitely longer
> > than
> > > reading it.
> > > So there is a prime after which is it worthwhile to store
> the larger
> > primes.
> >
> > The point is, it's not worth to store a table of _all primes_
> up to a
> > certain number. Sure, just reading a single prime can be
> fast, but if
> > you want to get all the primes in a reasonably large interval, it's
> > likely faster to generate them then to store them.
> >
> > Ambrus
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
