This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: altivec predicates
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Aldy Hernandez <aldyh at redhat dot com>, David Edelsohn <dje at watson dot ibm dot com>
- Cc: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Sun, 03 Feb 2002 12:52:57 -0800
- Subject: Re: altivec predicates
--On Sunday, February 03, 2002 02:35:30 PM +1100 Aldy Hernandez
<aldyh@redhat.com> wrote:
>>>>>> "David" == David Edelsohn <dje@watson.ibm.com> writes:
>
> > I guess I'm just being dense. I don't understand why you need to
> > explicitly create separate builtin builtin variants for each mode.
>
> No, I think you have a point.
I'm not sanguine about creating builtin functions that do not have
a well-defined type. From a language point of view, these builtins
should have a type (like "int ()(int, int)"). Otherwise, we're
asking for all kinds of problems. If the type involves variable
arguments, that's OK -- but then we should expect that the usual
promotion rules apply, so that, for example, you can never get
a float as one of the parameters.
To me, that means that you need a separate builtin for each mode.
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com