This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] more PR33928: DIY dataflow in fwprop, to enable it at -O1 and speed it up at -O2
On Fri, May 8, 2009 at 11:25 AM, Paolo Bonzini <bonzini@gnu.org> wrote:
>
>>> Err - if this patch doesn't make a difference what critical bug
>>> does it fix to support backporting to 4.4?
>>
>> What makes you think it does not make a difference?
>
> Well, I had placed in my list
>
> ? ?+ no code difference at -O2 before and after the patch (not tested
> ? ? ?except on the PR testcase), giving more confidence in backporting
>
> but...
>
>> Without this
>> patch, the compile time in fwprop for PR26854 and similar
>> dense-CFG-dense-DU/UD-web test cases is horrible.
>
> ... this was of course in the subject.
Ok, I indeed was confused about the "no code difference".
The patch is ok for trunk, let's see what issues pop up with
enabling fwprop at -O1 before backporting though.
Richard.
> Paolo
>