[Bug tree-optimization/46722] [4.6 Regression] Missed fma for x*x + y

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46722

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-30 14:38:51 UTC ---
Any reason why we prefer pow (x, 2) over x * x?  The latter is smaller tuple,
and we don't canonicalize x to pow (x, 1) either.