Re: Porposal: Floating-Point Options

[Geoffrey Keating <geoffk at geoffk dot org>]

Scott Robert Ladd <scott.ladd@coyotegulch.com> writes:

> To support different expectations, I suggest defining the following
> floating-point options for GCC. This is a conceptual overview; once
> there's a consensus the categories, I'll propose something more formal.
> 
> -ffp-correct
> 
> This option focuses code generation on mathematical correctness,
> portability, and consistency. No 80-bit long doubles, no fsin/fcos,
> making certain that comparison operators work within reason. Note that
> this option can only go so far in ensuring portability, given that not
> every system supports IEEE 754 floats.
> 
> -ffp-ieee754
> 
> To the best of our ability, assume and follow IEEE 754. Similar to the
> above, but assuming IEEE floats and doubles. Note that IEEE 754 has been
> undergoing some revision.
> 
> -ffp-balanced (default)
> 
> Balance correctness with speed, enabling performance optimizations that
> may reduce portability or consistency, but which do not alter the
> overall quality of results. Yeah, I know that's a bit fuzzy; formal
> definition of such an option depends on categorizing existing FP code
> generation (something I'm working on).
> 
> -ffp-damn-the-torpedoes-full-speed-ahead
> 
> Okay, maybe that should be something shorter, like -ffast-math or
> -ffp-fast. This option enables dangerous hardware intrinsics, and
> eschews all concerns about portability and consistency in the quest for
> speed. The behavior is likely to be the same as the current -ffast-math.

I think it would help to look at what we already have, first.  We have
-ffast-math, which has a bunch of different sub-options for people who
have done careful analysis of their code plus the
-funsafe-math-optimizations flag for people who don't care.

On many ports (like powerpc), the compiler is intended by default to
conform to C99 plus appendix F.  I think this is what you mean by
'-ffp-correct'.  This is already implemented by -fno-fast-math on
those ports.

So, what I think you want to do is to add another flag under
-ffast-math, perhaps called -fwiden-math, which specifically allows
the compiler to compute values in a wider mode (that would be 80-bit
x87 FP) even when ISO C doesn't allow it.  You would have this flag on
by default on x86.