This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Store motion related questions
- To: Daniel Berlin <dan at cgsoftware dot com>
- Subject: Re: Store motion related questions
- From: Stephen L Moshier <moshier at mediaone dot net>
- Date: Fri, 7 Sep 2001 14:39:56 -0400 (EDT)
- cc: gcc at gcc dot gnu dot org
- Reply-To: moshier at moshier dot ne dot mediaone dot net
> 1. Why would these failures only, and consistently, happen at -O3. Do
> we not try to simplify floating point values at -O2?
Right, O3 has the effect of turning functions in these test
cases into inline code, which gets optimized by the compiler
doing the whole calculation and just returning the answer
(not merely the answer of the arithmetic problem, just abort
or exit 0!) . In these cases the calculation involves evaluating
a constant expression and comparing with aonther constant.
Without the inlining, the constant expressions don't occur at O2.