Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site oblio.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!ihnp4!crsp!pesnta!hplabs!oblio!kent From: kent@oblio.UUCP (Kent Peacock) Newsgroups: net.lang.c Subject: Re: host 68K C cross-compiler sought Message-ID: <228@oblio.UUCP> Date: Fri, 18-Jan-85 20:05:04 EST Article-I.D.: oblio.228 Posted: Fri Jan 18 20:05:04 1985 Date-Received: Sun, 20-Jan-85 08:00:09 EST References: <7485@brl-tgr.ARPA> Organization: Counterpoint Computers Lines: 19 You might also want to consider the SGS 68000 compiler from AT&T. I spent some time awhile back tuning a version of the Stanford compiler (originally from MIT) to bring it close to the Green Hills compiler, which we also had for evaluation. I would say that the improvements brought the Stanford compiler from about 15% larger to less than 5% larger code. Most of the changes were obvious and took about 1 man-week to put in. When I changed companies, and compared the density of the tuned compiler with the SGS compiler, the SGS 68000 compiler came out about 2% better. The bottom line: the SGS compiler is pretty close to the Green Hills compiler in code density, and is written in C and based on PCC2. As I recall, the Green Hills compiler is written in PASCAL, which is good or bad, depending on your viewpoint (:-). The compiler I worked on is being used by Dialogic Systems for their own 68000-based product. I don't know if they are interested in distributing it. Kent Peacock Counterpoint Computers