Re: [PATCH] add statistics counting to postreload, copy-rename, and math-opts

[Richard Guenther <richard.guenther@gmail.com>]

On Tue, Apr 12, 2011 at 4:51 PM, Nathan Froyd <froydnj@codesourcery.com> wrote:
> On Tue, Apr 12, 2011 at 04:37:42PM +0200, Richard Guenther wrote:
>> On Tue, Apr 12, 2011 at 4:32 PM, Nathan Froyd <froydnj@codesourcery.com> wrote:
>> > Thanks.  I may go twiddle that patch to do something similar to mine and
>> > submit that.  Do you use your patch for checking that the same set of
>> > optimizations get performed, then?  I'm interested in using the
>> > statistics for identifying passes that don't buy us much across a wide
>> > variety of codebases.  (Suggestions for suitable ones welcome!)
>>
>> Yes, I used it exactly for that.  And also to verify that passes don't
>> do anything if replicated (well, for those that shouldn't at least).
>>
>> Don't expect any low-hanging fruit though ;)  I catched all of it already.
>>
>> Candidates are obviously SPEC and GCC itself.  I also use tramp3d
>> of course.  That said, even if a pass does nearly nothing we often
>> have testcases that need it ...
>
> True, but maybe those testcases should be adjusted--per-pass flags,
> rather than blindly assuming -O2 includes them.  And it's not clear to

It's easier to add things to GCC than to argue removing things ...

> me that the statistics_counter_event infrastructure really helps
> catching do-nothing passes, since it doesn't record stats that increment
> by zero...

Well, if the overall count is zero then nothing was done.

Richard.

> -Nathan
>