Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxl!ihnp4!zehntel!hplabs!sri-unix!gwyn@BRL-VLD.ARPA From: gwyn@BRL-VLD.ARPA Newsgroups: net.unix Subject: Re: Assembler vs C Message-ID: <159@sri-arpa.UUCP> Date: Wed, 25-Apr-84 06:33:18 EDT Article-I.D.: sri-arpa.159 Posted: Wed Apr 25 06:33:18 1984 Date-Received: Wed, 2-May-84 05:37:47 EDT Lines: 16 From: Doug Gwyn (VLD/VMB) About the only use of assembler on our many UNIX systems is by kernel workers or C run-time library maintainers in direct support of the software used by everybody else. If the UNIX System V manuals appear to dissuade programmers from using assembler, then that is a good sign. The ONLY times I resort to assembler rather than C are: -- function cannot be performed in C. There are very few cases where this is true; implementing setjmp/longjmp is a good example. -- C implementation is a bottleneck. One identifies such cases by profiling the code, then considering whether the improved speed is going to be worth the maintenance costs associated with use of assembler. Seldom is it really worthwhile to use assembler, all things considered.