Re: PATCH - [tree-ssa] regrouping of expression tree for single multiply add.

[Dale Johannesen <dalej at apple dot com>]

On Mar 25, 2004, at 3:23 PM, law@redhat.com wrote:
> > > > No.  Try mainline.
> > > And I think there's an excellent argument against this kind of
> > > transformation
> > > for the mainline as well, at least for FP types.
> >
> > What is the argument? If issue is correctness for corner cases, is 
> > there
> > additional gcc flag; besides -ffast-math, which need be checked for?
> Just watch the main development list -- it's being discussed 
> extensively
> under these threads:
>
> GCC options and floating point correctness
> (a + b) + c should be replaced by a + (b + c)
> GCC beaten by ICC in stupid trig test

Well, there's an argument against it, all right, but the view that it's 
excellent
appears to be a minority one.  It's clear that quite a few people find 
such
transformations useful.  Those of you who want standard-conforming 
behavior
don't have to use -ffast-math, after all.