This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


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

Re: Measuring gcc optimizations: are they well balanced?


At 15:36 10.01.00 , Andi Kleen wrote:
>On Mon, Jan 10, 2000 at 02:53:09PM +0100, Franz Sirl wrote:
> >
> > This seems to be a miscompilation on x86 only (?), I tried it with current
> > gcc-2_95-branch on PPC and the produced code looks and works fine. Andi
> > tried the current 2.95-branch on x86 and could still reproduce the bug.
>
>I just looked at it again, and gcc 2.95 current seems to generate
>correct code now. Maybe I just looked at the wrong assembler file.
>I'm sorry for the confusion. Jan, no need to do anything.

Just FYI, the patches that probably fixed this are:

Mon Dec 27 07:32:59 1999  Jeffrey A Law  (law@cygnus.com)

         Thu Sep 16 00:46:35 1999  Geoff Keating  <geoffk@cygnus.com>
         * loop.c (basic_induction_var): A non-integer variable which is
         being set by a paradoxical subreg is probably not a biv.

         1999-08-31 12:20 -0700 Zack Weinberg <zack@bitmover.com>
         * loop.c (check_dbra_loop): Check bl->biv->add_val is a
         CONST_INT before using its INTVAL.

My personal bet goes on Zack's patch :-). It fixed some strange behaviours 
on PPC too, that's why I told you to try the branch :-).

BTW, Andi, do you have any other serious 2.95.2 bugs in your queue?

Franz.


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