This is the mail archive of the gcc-patches@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]
Other format: [Raw text]

Re: [PATCH] Improve ifcombine (PR 52005)


On Tue, 24 Jul 2012, Andrew Pinski wrote:

On Tue, Jul 24, 2012 at 2:37 AM, Richard Guenther
<richard.guenther@gmail.com> wrote:
On Tue, Jul 24, 2012 at 10:25 AM, Marc Glisse <marc.glisse@inria.fr> wrote:
On Mon, 23 Jul 2012, Andrew Pinski wrote:

This patch improves ifcombine by doing two things.  First tries to see
if the then and else are swapped and handles that case.


Just a few hours after I posted:
http://gcc.gnu.org/ml/gcc-patches/2012-07/msg01160.html

Could you check that your patch works on my testcases? (and then we can add
them to the testsuite and forget about my patch)

I must say I like Marcs patch more ...

I agree it is better. Though I don't have time to test my testcases on your patch.

They won't work as is, because I just give up on bit tests in the inverted case. I guess replacing invert_tree_comparison with fold_build1(TRUTH_NOT_EXPR) at the beginning of the function might bring the patches closer (Andrew's patch seems to do more than mine, though I haven't studied it in details).


--
Marc Glisse


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