This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: IEEE 754 and fused-madd (was: Re: patch: rs6000 specific)
- From: Brad Lucier <lucier at math dot purdue dot edu>
- To: dje at watson dot ibm dot com (David Edelsohn)
- Cc: dalej at apple dot com (Dale Johannesen), shebs at apple dot com (Stan Shebs), lucier at math dot purdue dot edu, gcc-patches at gcc dot gnu dot org
- Date: Thu, 6 Dec 2001 19:45:56 -0500 (EST)
- Subject: Re: IEEE 754 and fused-madd (was: Re: patch: rs6000 specific)
> Second, this issue should not be considered two categories (IEEE
> 754 and non-IEEE 754) but three categories (round every intermediate
> expression, excess intermediate precision, and fast-and-loose).
How about these three categories: IEEE 754 without excess intermediate precision,
IEEE 754 with excess intermediate precision, and fast-and-loose.
> So, as far as I am concerned, FMA as implemented by GCC is allowed
> by default.
> Any transformations to utilize FMA which are allowed by GCC
> with the -ffath-math option are not allowed by default.
How about: Transformations that do not change the results of IEEE 754
arithmetic are allowed by default (presumably, some of these could help
the MD file to recognize opportunities for FMA use); transformations that
can possibly change the results of IEEE 754 arithmetic are not allowed by
default (even when FMA is allowed), but are allowed with
-funsafe-math-optimizations.
> Only expressions
> matching the various FMA instructions can be in the PowerPC MD.
This is the part of the discussion I don't understand; perhaps I don't
have to understand, or I might understand if given some examples.
Brad