From ... Path: supernews.google.com!sn-xit-02!supernews.com!isdnet!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!uio.no!Norway.EU.net!127.0.0.1!nobody From: Erik Naggum Newsgroups: comp.lang.lisp Subject: Re: Performance tuning Date: 05 Dec 2000 00:59:09 +0000 Organization: Naggum Software; vox: +47 800 35477; gsm: +47 93 256 360; fax: +47 93 270 868; http://naggum.no; http://naggum.net Lines: 23 Message-ID: <3184966749823376@naggum.net> References: <3184953833736032@naggum.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: oslo-nntp.eunet.no 975979022 27728 195.0.192.66 (5 Dec 2000 01:17:02 GMT) X-Complaints-To: newsmaster@eunet.no NNTP-Posting-Date: 5 Dec 2000 01:17:02 GMT mail-copies-to: never User-Agent: Gnus/5.0803 (Gnus v5.8.3) Emacs/20.7 Xref: supernews.google.com comp.lang.lisp:4705 * Greg Menke | I compiled the core functions in Personal Lispworks using strongest | optimization, which improved runtimes considerably- but it remained | extremely slow; minutes for a largish file compared to about a second | with the C++ version. I experienced similar performance problems with my first few shots. Today, a 640K file takes 190 ms for the md5sum utility to return, the new md5 function in Allegro CL takes as long with that file's contents as a string, but also conses equally much in static memory. My last version takes 130 ms and conses less than 100 bytes regardless of the length of the input string. It is not particularly portable. It is also intended to be used as a sink stream so writing to the stream computes the md5sum, retrievable upon closing the stream. This is not quite as fast as I want it to be, but I expect this to change when I have reimplemented it using the new Allegro CL streams concepts. #:Erik -- "When you are having a bad day and it seems like everybody is trying to piss you off, remember that it takes 42 muscles to produce a frown, but only 4 muscles to work the trigger of a good sniper rifle." -- Unknown