What is acceptable for -ffast-math? A numerical viewpoint

[Ross Smith]

dewar@gnat.com wrote:
> 
> <<Having optimization change program results is a trickier case.
> 
> But *all* the "optimizations" you suggest *do* change program results!

Not enough to matter. This is the point you and Gabriel seem incapable
of understanding.

Who decides what's "enough to matter"? _We_ do. They're our programs.

Stop telling us what we're allowed to want.

-- 
Ross Smith <ross.s@ihug.co.nz> The Internet Group, Auckland, New Zealand
========================================================================
"Unix has always lurked provocatively in the background of the operating
system wars, like the Russian Army."                  -- Neal Stephenson