Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84 SMI; site sun.uucp Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!mtuxo!mtunh!mtung!mtunf!ariel!vax135!timeinc!phri!pesnta!amd!amdcad!decwrl!sun!jfarrell From: jfarrell@sun.uucp (Jerry Farrell) Newsgroups: net.lang.c Subject: Re: more questions about efficient C code Message-ID: <2367@sun.uucp> Date: Sun, 7-Jul-85 00:01:12 EDT Article-I.D.: sun.2367 Posted: Sun Jul 7 00:01:12 1985 Date-Received: Wed, 3-Jul-85 08:32:09 EDT References: <474@crystal.UUCP> <420@enmasse.UUCP> Reply-To: jfarrell@sun.UUCP (Jerry Farrell) Distribution: net Organization: Sun Microsystems, Inc. Lines: 20 In article <420@enmasse.UUCP> mhs@enmasse.UUCP (Mike Schloss) writes: >> I am currently modifying C code, written by someone else, >> that is incredibly terse. It's paramount that the code >> be fast, so I presume the code is terse for speed. I'm >> now curious about a few things: > >Sorry but it all depends on your particular compiler >implementation and the machine you are running on. >There are two ways to find the answers to your questions. >One is to write a simple test program and use the '-S' >flag of the compiler. This will produce assembly code >that you can examine and see what intructions are being >generated for any particular expression. Unfortunately, this option is not likely to show you the optimized code you would get from running the compiler with the -O flag. To really see what's being generated, compile it, and then look at the object file with something like foo,24?ai in adb.