This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: support of 128 bit long real on linux for ia32
- From: Jakub Jelinek <jakub at redhat dot com>
- To: Stephen L Moshier <steve at moshier dot net>
- Cc: Winfrid dot Tschiedel at hpc dot fujitsu-siemens dot com, gcc at gcc dot gnu dot org
- Date: Wed, 13 Feb 2002 08:41:32 -0500
- Subject: Re: support of 128 bit long real on linux for ia32
- References: <Pine.LNX.4.44.0202130740500.23743-100000@moshier.net>
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
On Wed, Feb 13, 2002 at 08:07:43AM -0500, Stephen L Moshier wrote:
>
> > Do plan to enhance your compiler in order to get
> > "long double" in C compatible with real(16) in fortran ?
>
> It was a straightforward job to implement the "long double"
> type in i386 gcc as 128-bit soft-float TFmode with 113-bit precision.
> I might still have the patches for that (though they would be about
> eight years old now).
>
> You cannot easily have both 80-bit and 128-bit support in gcc at the
> same time, however. I think you might have to implement
> a "long long double" which probably no one has done yet.
I don't think you need long long double, double __attribute__((mode(TF)));
is IMHO the way to go.
Just IMHO we need separate TXFmode, which would be the current TFmode in
with INTEL_EXTENDED_IEEE_FORMAT != 0 (and TFmode would be current TFmode
in non-intel format).
Jakub