Path: utzoo!attcan!uunet!mcsun!ukc!dcl-cs!aber-cs!athene!pcg From: pcg@cs.aber.ac.uk (Piercarlo Grandi) Newsgroups: comp.lang.misc Subject: Re: Aggressive optimization Message-ID: Date: 8 Nov 90 16:29:45 GMT References: <2060@aber-cs.UUCP> <65592@lanl.gov> <2677@l.cc.purdue.edu> <12175@ganymede.inmos.co.uk> <123@garth.UUCP> Sender: pcg@aber-cs.UUCP Organization: Coleg Prifysgol Cymru Lines: 48 Nntp-Posting-Host: odin In-reply-to: smryan@garth.UUCP's message of 6 Nov 90 20:12:12 GMT On 6 Nov 90 20:12:12 GMT, smryan@garth.UUCP (Steven Ryan) said: smryan> If it's optional and you don't want, don't use. If you can buy a smryan> compiler without an optimiser (which is easy), do so. If nobody smryan> wants an optimiser, nobody buys them. If somebody wants one, why smryan> not let them buy it? Oh, this is sociology -- the issue is not whether there is a *market* for optimizers or not -- it is whether the market should exist, and which needs it should satisfy. People who buy optimizers do so not on the basis of adverts associating them with scantily clad gorgeous females or (even more alluring to some people :->) 64 bit superscalar CPUs. They (should) do so on the basis of a cost effectiveness argument, and we have been debating this. The argument is actually twofold: as a user, should I buy an optimizer? as a developer, should I invest in building an optimizer? The second problem is also very important -- the computer industry is largely driven by the supply, not the demand side, quantitatively and qualitatively. Users look to implementors for a sense of direction. The decisions of implementors will be often followed without questioning, because users do not have the experience or insight and more or less know it. IMNHO the main reason for which the computer industry is driven by the supply side is that talent is very scarce, and for this reason not all possible alternatives worth exploring can be explored. The problem, bfore considering market demand, can also be expressed thus: should David Chase and Jim Giles and Preston Briggs and Hank Dietz (just to mention some frequent contributors to this sort of discussion) spend their time devising ever more ingenious and complicated and hairy code generator level hacks, or would their research and development efforts be better spent on clever compilation of higher level languages and source level programmer's assistants? Are there enough compiler experts to explore both possibilities? If not, which possibility looks the best winner? Is aggressive optimization of low level languages a technological blind alley like autocoders? Inquiring minds want to know... Not simple questions, and I am presenting here an unfashionable but hopefully amusing set of answers to them, which should not be ignored. -- Piercarlo Grandi | ARPA: pcg%uk.ac.aber.cs@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcsun!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk