Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!killer!ames!pasteur!ucbvax!decwrl!pyramid!cbmvax!snark!eric From: eric@snark.UUCP (Eric S. Raymond) Newsgroups: comp.arch Subject: Re: Compiler complexity (was: VAX Always Uses Fewer Instructions) Summary: You forgot the RISC crowd's favorite horror story Message-ID:Date: 23 Jun 88 11:02:36 GMT References: <6921@cit-vax.caltech.edu> <28200161@urbsdc> <10595@sol.arpa> <20338@beta.lanl.gov> <1117@ima.isc.com> <1127@ima.isc.com> Organization: Willam Claude Dukenfield Discordian Cabal Lines: 18 In article <1127@ima.isc.com>, samples@dougfir.Berkeley.EDU (A. Dain Samples) writes: >A slight correction needs to be made here: there is only one instance >that I know of where a complex instruction on a CISC architecture turned >out to run more slowly than the same operation coded with simpler >instructions ON THE SAME MACHINE. And the moderator clarifies: >[As dmr noted in his recent comp.arch note, it was the 780's calling >instructions.] Eh? You both forgot the RISC crowd's favorite horror story -- the VAX POLY instruction for evaluating a polynomial in x in one swell foop, given x and the coefficients. Turns out this is always slower than a loop using simpler instructions. -- Eric S. Raymond (the mad mastermind of TMN-Netnews) UUCP: {{uunet,rutgers,ihnp4}!cbmvax,rutgers!vu-vlsi,att}!snark!eric Post: 22 South Warren Avenue, Malvern, PA 19355 Phone: (215)-296-5718