Minimum floating-point requirements

[Joseph S. Myers]

On Mon, 17 Feb 2014, Steve Munroe wrote:

> And why do I keep find packages with -ffloat-store forced on? This is bad
> because PowerPC always had IEEE compliant double and this really hurts
> performance.

I think -ffloat-store should be a no-op on platforms without excess 
precision (and for that matter on platforms where the FLT_EVAL_METHOD 
definition has excess precision but that's done for consistency with 
mistaken float_t / double_t definitions in glibc, rather than because the 
platform actually lacks native float/double arithmetic, i.e. S/390).  
I.e., GCC should be changed to optimize this.

-- 
Joseph S. Myers
joseph@codesourcery.com