Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!iuvax!jec From: jec@iuvax.cs.indiana.edu (James E. Conley) Newsgroups: comp.arch Subject: Re: BENCHMARKS AND LIPS Message-ID: <15488@iuvax.cs.indiana.edu> Date: 2 Dec 88 17:05:15 GMT References: <1740MLWLG@CUNYVM> <746@quintus.UUCP> <595@mqcomp.oz> <1219@cps3xx.UUCP> <429@babbage.acc.virginia.edu> Reply-To: jec@iuvax.UUCP (James E. Conley) Organization: Indiana University CSCI, Bloomington Lines: 20 MIPS and LIPS have a similar problem-- that you may not be counting the same sorts of instructions or logical inferences. The VAX instruction to add two integers and the MIPS instruction really don't do the same thing and similarly a logical inference has the same sort of problem-- even if you are dealing with the same system (Prolog in this case). This gets worse if you start comparing other systems which do their inferences in a different manner (production systems like YAPS and OPS5 come to mind). As usual, benchmarking using LIPS will get you a number with no solid meaning (as with MIPS). III Usenet: iuvax!jec UUU I UUU ARPANet: jec@iuvax.cs.indiana.edu U I U Phone: (812) 855-7729 U I U U.S. Mail: James E. Conley U I U Indiana University UUUIUUU Dept. of Computer Science I 004 Lindley Hall III Bloomington, IN. 47405