[Bug tree-optimization/48738] pow() fails to produce (some) subnormalized numbers with integer exponents

[wschmidt at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48738

--- Comment #6 from William J. Schmidt <wschmidt at gcc dot gnu.org> 2011-05-25 12:15:46 UTC ---
(In reply to comment #5)
> I don't think this is a library issue. Maybe Bill is interested / has an
> opinion..

Just to agree that it is WAD.  The proposed solution just moves the values of x
and m for which the rounding will cause edge-case anomalies.  Either way, the
compounded rounding error means the two versions of pow will never produce
identical results.  When precision is important, use the real-valued version.