This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: possible fix for PR 32349
On 6/15/07, Kenneth Zadeck <zadeck@naturalbridge.com> wrote:
Janis Johnson wrote:
> On Fri, 2007-06-15 at 18:43 -0400, Kenneth Zadeck wrote:
>
>> I have only tested this with the testcase.
>>
>
> With the patch I get no failures from SPEC CPU2000 on
> powerpc64-linux compiled with "-O2 -fmodulo-sched" and
> either -m32 or -m64. Without the patch there were 3
> failures for -m32, 6 for -m64.
>
> Janis
>
>
These are a distressingly low numbers. i would have expected this bug
to get a lot more hits on the fortran benchmarks.
SPEC CPU2000 fp has 10 fortran programs.
This means more than half of them failed with -m64.
That doesn't sound ""distressingly low",
although if you expect it to fail whenever
modulo scheduler worked, certainly this looks low
- probably that's more of an issue for the modulo scheduler than anything else.
--
#pragma ident "Seongbae Park, compiler, http://seongbae.blogspot.com"