Skip to Content.
Sympa Menu

sm-discuss - Re: [SM-Discuss] Smallest Source Mage

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: Jon Svendsen <jon-sven AT frisurf.no>
  • To: "M.L." <m.l AT compaqnet.fr>
  • Cc: sm-discuss <sm-discuss AT lists.ibiblio.org>
  • Subject: Re: [SM-Discuss] Smallest Source Mage
  • Date: 29 Jul 2002 02:19:01 +0200

On Sun, 2002-07-28 at 16:30, M.L. wrote:
> > No. Each new family of CPUs introduce new instructions that weren't
> > present in older familys. If you compile binaries using these
> > instructions, and then try to run them on a system with a lower class
> Fine... I'm a bit surprised here, since I tought that this kind of things
> should not happen.

You actually choose whether you want it to happen or not. With gcc, the
-mcpu option controls which processor family to optimize for
(instruction scheduling etc), and the -march option controls which
instruction set to use. By setting march higher than i386 you explicitly
say that "I know that I won't be able to run these binaries on lower
class boxes, so I hope you can make them a little faster in return".
This is why you can get your box running faster with source based
distributions than with binary. redhat compiles their binaries
-march=i386 -mcpu=i686, mandrake i think uses -march=i586 -mcpu=i686






Archive powered by MHonArc 2.6.24.

Top of Page