This is the mail archive of the crossgcc@sourceware.org mailing list for the crossgcc project.

See the CrossGCC FAQ for lots more information.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Possibly OT: GCC 2.96 on PPC for VxWorks 5.5


Paul Smith schrieb:
Well, the compile lines I use to generate the .s files on both systems
are byte-for-byte identical, so it's not the command-line flags
(optimization levels, -f arch settings, etc.)  I've also looked at the
ccppc -v output and compared those: no differences.

The spec file that was installed for Linux was very different than the
one for Solaris.  However, I did try copying the Solaris spec file over
into the Linux compiler installation and that didn't help either.

I have not had to sleuth through spec files, or even getting gcc to compile for a 'new target',
in quite some time... perhaps 10 years. I've also sworn off WRS for about as long as well.


However, back when I was trying to understand gcc wyrdnesses, what was listed in the
spec file was important, as that's were any 'default' specs are used, and not necessarily in
the expected order, or conditions. In the code example you gave, my 'first guess' would be
that there's an optimization in the one case to simplfy the code to a single multiply and store, but
it could also be used because in the 'WRS' spec, a simpler machine comes first in the default selection
order.


Just as an aside, if you have the compiler to the point were it generates so close to nearly identical
code, what is the 'problem'?


John Clark


-- For unsubscribe information see http://sourceware.org/lists.html#faq


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]