Athlon optimizations

Ainsley Pereira marble at
Tue Nov 21 11:29:55 PST 2000

In message <20001120231917.D9DEC36E60 at>
          scott thomason <scott at> wrote:

> I second your suspicion. Using -malign-double will certainly break
> some cherished expectations that one var will directly follow another
> in memory. You may say that would be a bad programming practice, and
> you would be right, but there are some libraries, for example,
> libtiff, that need to have one follow another in order to comply with
> binary format standards.

*Sigh* if only they'd standardise #pragma pack(1) or __attribute__((packed))
(disclaimer: the gcc version may be wrong; I have to develop with msvc at work,
so I know the pragma...)

> Out of curiousity, were you using AGCC, pgcc, or gcc?

Plain gcc. It was a 'more or less LFS but with whatever versions of things I
have handy'. I'll go through and upgrade everything to the latest versions
soon, and give AGCC or pgcc a go then:)


   / Ainsley Pereira       /   aka Marble on Snowplains   /
  /                       / telnet:// /

Unsubscribe: send email to lfs-apps-request at
and put unsubscribe in the subject header of the message

More information about the blfs-support mailing list