[Bug rtl-optimization/42863] [4.7/4.8/4.9 Regression] with auto increments

law at redhat dot com gcc-bugzilla@gcc.gnu.org
Mon Feb 10 23:15:00 GMT 2014


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42863

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |law at redhat dot com
         Resolution|---                         |FIXED

--- Comment #9 from Jeffrey A. Law <law at redhat dot com> ---
AFAICT the current trunk generates code that is as good as or better than
gcc-4.4.

The redundant moves/tests are gone.  And the load of "dst" out of the stack is
deferred until we know we need it.   There's a tiny bit of register shuffling
in one arm of a conditional, but it's offset by avoid it in another
conditional.

Overall, the trunk code is IMHO as good as or better than gcc-4.4. 

There are no plans to identify what changes caused the improved code
generation.  Obviously with no plans to identify the changes, there are no
plans to backport the changes to older release branches.



More information about the Gcc-bugs mailing list