Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: Notesfiles $Revision: 1.6.2.17 $; site ea.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxj!ihnp4!inuxc!pur-ee!uiucdcs!ea!mwm From: mwm@ea.UUCP Newsgroups: net.ai Subject: Re: Compiled versus interpreted LISP Message-ID: <500003@ea.UUCP> Date: Wed, 24-Oct-84 21:09:00 EST Article-I.D.: ea.500003 Posted: Wed Oct 24 21:09:00 1984 Date-Received: Sun, 28-Oct-84 06:54:55 EST References: <63@water.UUCP> Lines: 22 Nf-ID: #R:water:-6300:ea:500003:000:773 Nf-From: ea!mwm Oct 24 15:09:00 1984 I have in front of me the UOLISP manual. I quote from it: The UOLISP philosophy places greater emphasis on compiled code since these functions take 1/2 to 1/3 the space of interpreted ones as well as running 20 to 75 times faster. I just got my copy of UOLISP, and haven't had time to verify this. I plan on doing so in the near future. Among the reference I also find: Marti, J., 'An Optimizing Compiler for LISP for the Z80', SIGSMALL 1982. The z80 is a special case for optimization, as it's addressing modes don't help with stack fetching except under special conditions. See the October 1984 Dr. Dobb's Journal for an optimization scheme that can help a z80/8080, but is totally worthless for a reasonable processor (6809, 68000, VAX, DEC-10, etc.)