[Bug middle-end/39838] [5/6 regression] unoptimal code for two simple loops

[law at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=39838

Jeffrey A. Law <law at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[5/6/7 regression]          |[5/6 regression] unoptimal
                   |unoptimal code for two      |code for two simple loops
                   |simple loops                |

--- Comment #25 from Jeffrey A. Law <law at redhat dot com> ---
I don't think the codesize going up with -O2 is a significant issue here.  It
looks like the regression with -Os is fixed.  We can also see that we're doing
less memory traffic.

I think we should mark this as resolved for gcc-7.