Date: Wed, 24 Nov 2004 08:24:28 +0100 From: Ingmar Gebert <ingmar.gebert@uni-rostock.de> To: FUJISHIMA Satsuki <sf@FreeBSD.org> Cc: asa@gascom.ru Subject: Re: missing libmad optimization, diff included Message-ID: <001201c4d1f6$a3821db0$6a111e8b@Pequod> References: <000401c4d196$ae1dbaa0$d2111e8b@Pequod> <867joc6js7.wl%sf@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
In fact I did measure the performance -- it shaves around 7% off the cpu hit. Yes, that's on a low spec machine (Pentium 200 MMX) and probably negligable on today's machines. But the way I see it: the knob doesn't hurt anybody, as the default behaviour is kept, and the user gets more options. Also, the mpglib alternative doesn't help as the software I'm using (musicpd) links to libmad. ----- Original Message ----- From: "FUJISHIMA Satsuki" <sf@FreeBSD.org> To: "Ingmar Gebert" <ingmar.gebert@uni-rostock.de> Cc: <asa@gascom.ru>; <ports@FreeBSD.org> Sent: Wednesday, November 24, 2004 3:52 AM Subject: Re: missing libmad optimization, diff included > Did you measured their performance? For me --with-accuracy outperforms > --with-speed so I chose accuracy (3.5 years ago). > > libmad is for accuracy. If you want speed, mpglib runs way faster. :) > > I don't object at all if you really want that knob though. > > At Tue, 23 Nov 2004 20:57:18 +0100, > Ingmar Gebert wrote: >> >> I miss a knob for compiling libmad with speed optimizations, >> which is offered during configuration. See the diff below for the simple >> changes. >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?001201c4d1f6$a3821db0$6a111e8b>