From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/3412 Path: news.gmane.org!not-for-mail From: Rich Felker Newsgroups: gmane.linux.lib.musl.general Subject: musl size breakdown Date: Sun, 16 Jun 2013 02:25:07 -0400 Message-ID: <20130616062507.GA19067@brightrain.aerifal.cx> Reply-To: musl@lists.openwall.com NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: ger.gmane.org 1371363921 22920 80.91.229.3 (16 Jun 2013 06:25:21 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Sun, 16 Jun 2013 06:25:21 +0000 (UTC) To: musl@lists.openwall.com Original-X-From: musl-return-3416-gllmg-musl=m.gmane.org@lists.openwall.com Sun Jun 16 08:25:22 2013 Return-path: Envelope-to: gllmg-musl@plane.gmane.org Original-Received: from mother.openwall.net ([195.42.179.200]) by plane.gmane.org with smtp (Exim 4.69) (envelope-from ) id 1Uo6Ou-0002PJ-FM for gllmg-musl@plane.gmane.org; Sun, 16 Jun 2013 08:25:20 +0200 Original-Received: (qmail 17590 invoked by uid 550); 16 Jun 2013 06:25:20 -0000 Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: Original-Received: (qmail 17582 invoked from network); 16 Jun 2013 06:25:19 -0000 Content-Disposition: inline User-Agent: Mutt/1.5.21 (2010-09-15) Xref: news.gmane.org gmane.linux.lib.musl.general:3412 Archived-At: Completely un-rigorous at the moment, but I thought these figures may be interesting. All size contributions are approximations based on my i386 build with -O3 and the size -t command, converted to percentages of the whole library size via quick mental math. 20% math (including complex) 15% iconv 10% stdio 10% regex/glob 10% crypt 5% network (esp. dns) 4% unicode tables (ctype) 3% string functions 3% threads 2% numeric parser/converters 2% time handling 1% malloc 1% unistd That leaves about 14% unaccounted for, presumably in lots of small stuff that just adds up. It would be interesting to see how these percentages compare with other implementations, and how users react to them and whether they're perceived as justified. With the further iconv features that are planned for the future, as well as corrections to the complex math code, math and iconv will have an interesting battle ahead to claim the size title. Math has an even greater chance of keeping the title if we consider non-i387 archs where most math functions involve a great deal of code rather than tiny asm implementations. Rich