home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!mcsun!sun4nl!spider!loki!bjl
- From: bjl@loki.pttrnl.nl (Ben Lippolt)
- Subject: Re: tlA: /bin/arch solved.
- Message-ID: <bjl.712133053@freyr>
- Sender: usenet@spider.research.ptt.nl (USEnet News)
- Nntp-Posting-Host: loki
- Reply-To: B.J.Lippolt@research.ptt.nl
- Organization: PTT Research
- References: <1992Jul23.170723.15247@klaava.Helsinki.FI> <bjl.711923672@freyr> <4120@inca.comlab.ox.ac.uk>
- Date: Sun, 26 Jul 1992 06:44:13 GMT
- Lines: 19
-
- as@prg.ox.ac.uk (Andrew Stevens) writes:
- >In article <bjl.711923672@freyr> B.J.Lippolt@research.ptt.nl writes:
- >>Is there a way to distinguish between a 386 and a 486? I have a 486,
- >>but 'uname -m' says 'i386'. But when I compile with 'gcc -m486' I'll
- >>get code which doesn't run on a 386.
-
- >Surely this is not the case. As I understood it the user-level instruction
- >sets of the 386 and 486 are identical. The -m486 flag just tweaks
- >code generation to use instruction sequences better suited to the 486's
- >instruction timings instead of sequences best suited to 386 instruction
- >timings.
-
- Thanks for your explanation. I was under the impression that the 486 had
- some extra instructions which the 386 didn't had.
- Does this mean that if I compile with -m486 and run the code on a 386 it
- will be slightly less efficient than without -m486?
-
-
- Ben Lippolt
-