home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.compilers
- Path: sparky!uunet!world!iecc!compilers-sender
- From: henry@zoo.toronto.edu (Henry Spencer)
- Subject: Re: static estimation of conditional branches?
- Reply-To: henry@zoo.toronto.edu (Henry Spencer)
- Organization: U of Toronto Zoology
- Date: Fri, 11 Dec 1992 17:41:18 GMT
- Approved: compilers@iecc.cambridge.ma.us
- Message-ID: <92-12-047@comp.compilers>
- Keywords: optimize, design
- References: <92-12-029@comp.compilers> <92-12-046@comp.compilers>
- Sender: compilers-sender@iecc.cambridge.ma.us
- Lines: 21
-
- hrubin@pop.stat.purdue.edu (Herman Rubin) writes:
- >... the programmer should have some way of communicating this, and other,
- >frequency considerations to the compiler. As far as I know, this has not
- >been done since the "FREQUENCY" statement in fairly early Fortran.
- >[As I recall, Fortran II dropped FREQUENCY because it was infrequently
- >used and made little difference. I've heard that it may even have been
- >implemented backwards and nobody noticed. -John]
-
- The other problem that occurs with such facilities is that programmer
- intuition is notoriously unreliable about such things.
-
- Now, if you amend Herman's statement to "the *profiler* should have some
- way of communicating this...", I'd agree... and observe that there are
- already compilers that will accept profiler data and exploit it for
- optimization. I don't know whether they do this particular optimization,
- but I wouldn't be surprised.
- --
- Henry Spencer @ U of Toronto Zoology, henry@zoo.toronto.edu utzoo!henry
- --
- Send compilers articles to compilers@iecc.cambridge.ma.us or
- {ima | spdcc | world}!iecc!compilers. Meta-mail to compilers-request.
-