Date: Thu, 26 Feb 1998 09:33:56 +0100 From: Eivind Eklund <eivind@yes.no> To: Anatoly Vorobey <mellon@pobox.com>, hackers@FreeBSD.ORG Subject: Re: RE: New utilities: factor(1) and wid(1)? Message-ID: <19980226093356.27757@follo.net> In-Reply-To: <19980226002846.05689@techunix.technion.ac.il>; from Anatoly Vorobey on Thu, Feb 26, 1998 at 12:28:46AM %2B0200 References: <01BD420C.2FCDD020.meuston@jmrodgers.com> <19980226002846.05689@techunix.technion.ac.il>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Feb 26, 1998 at 12:28:46AM +0200, Anatoly Vorobey wrote: > It won't, at least not easily. It keeps a fixed precomputed table > of all primes up to 2^16 to do a simple sieve (in > /usr/src/games/primes/pr_tbl.c, also used by primes(1)); would be, > err, an interesting exercize to change that to a fixed precomputed > table of all primes up to 2^32 and watch the executable size go up, > up, up into tens of megabytes and beyond... > > It's time for to rewrite them both to use a more modern method > of factoring, I guess. It's been some 2500 years or so; ole' good > Eratosthenes could use some rest :) If somebody need a pure prime-generator, I've got one lying around somewhere. A full sieve, no pre-computed tables. Not top-notch for really large sieving, but OK for <32 bits at least (and could AFAIR scale beyond that; I think I did it as a 2-layer sieve to be able to scale to infinity. Still won't be the fastest method for factoring, though.) License negotiatable (BSD, public domain, GPL - anything you want, as long as it doesn't involve me paying anybody else money at a later date ;-) Eivind. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980226093356.27757>