Re: gcc-2.7.2's optimizations

Bradley Ward Allen (ulmo@q.net)
23 Oct 1996 13:12:51 -0400

For a compiler which was originally touted to reduce an extremely
complex C file full of mathematical equations down to nothing if it
always had the same result quite a few years (a decade?) ago, I
have always *assumed* that it could handle a case such as that.

Also note gcc 2.7.2.1 has been out for a while now, which has the
long-awaited strength-reduce bug removed (I'm not sure if the
optimization was (a) removed, (b) made correct and optimal, or (c)
fixed but not made optimal). I use gcc 2.7.2.1 to compile everything,
and just use -O6 without a problem.