This is the mail archive of the
mailing list for the GCC project.
Re: combine permutations in gimple
On Wed, Aug 15, 2012 at 1:22 PM, Marc Glisse <firstname.lastname@example.org> wrote:
> On Mon, 13 Aug 2012, Marc Glisse wrote:
>> I'll give it a few more days for the conversation to settle, so I know
>> what I should do between:
>> - the barely modified patch you accepted,
>> - the check asked by Jakub,
>> - the restriction to identity that prevents any regression (well...),
>> - something else?
> I didn't realize the conversation would go quiet immediatly...
> Is someone willing to take the responsibility of telling me which approach
> is right? I can add Jakub's checks (though I am not sure how I would test
> them), but I would rather not do it if the conclusion is that they are
> either unnecessary (original patch is ok) or insufficient (don't avoid
> Ramana's regressions). I can do the very safe option 3 (only combine
> permutations when the result is the identity permutation), but I first want
> to make sure the more general thing is bad.
> (sorry, it's my first patch that gets conflicting answers...)
Well, we're waiting for someone to break the tie ... I'd go with the original
patch, improving the backends where necessary.
> Marc Glisse