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:
Hi all;
As of now, most of my .o's are building identically.  However, there're
some objects which don't.  After much investigation I've discovered that
for some .c files, the Linux version of the compiler generates this
assembly code:

       mulli 9,0,100
       stw 9,28(31)

while the Solaris version generates this assembly code:

       mr 11,0
       slwi 9,11,1
       add 9,9,0
       slwi 11,9,5
       add 9,9,11
       add 0,9,0
       stw 0,28(31)

Having spent thousands of dollars in psychoanalysis to irradicate any conscious or subconscious
material involved with any WRS product...


I would conjecture that the above is due to some form of option that selects a 'mul' over a shift
and add implementation. That could be based on machine type (ie which PPC implementation)
or optimizer settings, again conditioned on machine type, that the new compiler defaults to.
(PPC things for me are pretty ancient, but arn't there versions which don't have a multiply?)


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]