Mailinglist Archive: opensuse-bugs (8956 mails)

< Previous Next >
[Bug 186074] Optimize for i686
  • From: bugzilla_noreply@xxxxxxxxxx
  • Date: Tue, 10 Oct 2006 13:53:41 -0600 (MDT)
  • Message-id: <20061010195341.A033025C88E@xxxxxxxxxxxxxxxxxxxxxx>
https://bugzilla.novell.com/show_bug.cgi?id=186074





------- Comment #25 from to_jest_moj_mail@xxxxx 2006-10-10 13:53 MST -------
Well... Yes.

Guys from opensuse.org SUPER project have already tried and tested this, and
their benchmarks (dunno what kind of benchmarks, but I'm sure they knew what
they were doing) prove that it gives a performance improvement. I'd ommit that
O3 flag tho'.

Besides Simon Strandman wrote:

"...I did some testing with CFLAGS and lame. Lame where 17 seconds faster
encoding
a mp3 when compiled with "-march=i686 -mtune=generic" instead of "-march=i586
-mtune=i686". There are probably a lot of other packages that would give
similar results. Should we really miss this extra performance just to keep
compatibility with old i586 pentiums when 99,9% of the systems are various
i686s? Nobody uses i586 pentiums anymore anyway. "

His explanations and arguments are pretty good too.

Why are we optimizing stuff for machines that won't even be able to run it at
decent speed, no matter how much you tweak the software...

Besides, would it be such a big deal to change two small flags in gcc options
and then provide a repo & release with i686 packages for the users? This would
leave a room for choice. VIA Epia, or P266 user could download an i586 build
and someone with let's say AMD Athlon, or Pentium 4 could get the i686 version,
right?


--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

< Previous Next >