Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!apple!versatc!mips!mash From: mash@mips.COM (John Mashey) Newsgroups: comp.arch Subject: Re: 80486 vs. 68040 code size Message-ID: <18944@winchester.mips.COM> Date: 7 May 89 01:26:51 GMT References: <902@aber-cs.UUCP> Reply-To: mash@mips.COM (John Mashey) Distribution: eunet,world Organization: MIPS Computer Systems, Sunnyvale, CA Lines: 21 In article <902@aber-cs.UUCP> pcg@cs.aber.ac.uk (Piercarlo Grandi) writes: .... >The 68020 has more registers than the 386, but this ought to count for >little, as according to literature the knee of the curve for C is before or >at 4 registers (important note: for typical Unix utilities, e.g. excluding >floating point) because of the average simplicity of C expressions. Could you post the specific reference, with detail about what the circumstances that led to this conclusion? 1) It is very easy to generate such a conclusion for a specific machine & compiler combination, such that the conclusion is totally irrelevant to other combinations. For example, I once saw a statement that data on a PDP-10 said that the compilers seldom used more than 15 registers.....well, a PDP-10 has 16 registers.... 2) Good optimizing compilers make the number go up; in any case, the number 4 seems a bit low for a generic number. -- -john mashey DISCLAIMER: UUCP: {ames,decwrl,prls,pyramid}!mips!mash OR mash@mips.com DDD: 408-991-0253 or 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086