Re: -ffast-math and gfortran

[Toon Moene <toon at moene dot indiv dot nluug dot nl>]

I wrote:

> The most constructive criticism of -ffast-math (which I use exclusively,
> because I know what I'm doing, and repair damage as it appears),
> is to write documentation as to what changes -ffast-math allows.

> I hope I have time this weekend.

Sigh ... never promise to do something before you estimated the time
it takes to complete the task.

Especially -funsafe-math-optimizations nowadays does so much more than
a couple of years ago, that it is a really a lot of work to describe
all transformations allowed + the corner cases where the transformations
deliver obviously invalid results.

I'm sorry, I pass.

If you want to get a flavor of the number of places where the flag is
consulted, a grep -l unsafe_math *.c in the gcc/ directory points to
14 files ...  builtins.c uses it in around two dozen places, for instance.

Sorry,

-- 
Toon Moene - e-mail: toon@moene.indiv.nluug.nl - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/
Gfortran At Work: http://mysite.verizon.net/serveall/moene.pdf