This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Patch: Add BID decimal floating point format


On Wed, Feb 28, 2007 at 05:15:51PM -0500, Meissner, Michael wrote:
> > -----Original Message-----
> > From: Florian Weimer [mailto:fw@deneb.enyo.de]
> > Sent: Tuesday, February 27, 2007 4:20 PM
> > To: Meissner, Michael
> > Cc: gcc-patches@gcc.gnu.org; Harle, Christophe; H. J. Lu;
> bje@au1.ibm.com;
> > Janis Johnson
> > Subject: Re: Patch: Add BID decimal floating point format
> > 
> > * Michael Meissner:
> > 
> > > Are there any objections or comments to patch?
> > 
> >  @item --enable-decimal-float
> > +@itemx --enable-decimal-float=yes
> > +@itemx --enable-decimal-float=no
> > +@itemx --enable-decimal-float=bid
> > +@itemx --enable-decimal-float=dpd
> >  @itemx --disable-decimal-float
> >  Enable (or disable) support for the C decimal floating point
> > -extension.  This is enabled by default only on PowerPC GNU/Linux
> > -systems.  Other systems may also support it, but require the user to
> > -specifically enable it.
> > +extension.  This is enabled by default only on PowerPC, i386, and
> > +x86_64 GNU/Linux systems.  Other systems may also support it, but
> > +require the user to specifically enable it.  You can optionally
> > +control which decimal floating point format is used (either @var{bid}
> > +or @var{dpd}).  The @var{bid} format is default on i386 and x86_64
> > +systems, and the @var{dpd} format is default on PowerPC systems.
> > 
> > @samp{bid} and @samp{dpd} instead of @var{bid} and @var{dpd} (@var{}
> > is intended for metasyntactic variables only and is rendered in
> > CAPITAL LETTERS in plain text).
> > 
> > Are DPD and BID universal terms?  If not, it might make sense to
> > reference a definition (or provide one).
> > 
> 
> I'm attaching my rework of the documentation file.

The support in GCC is based on ISO/EIC WDTR 24732, which in turn is
based on IEEE 754r.

It doesn't make sense to enable decimal-float for targets other than
the ones mentioned because it requires support in the back end.

We shouldn't enable support by default because decNumber code in GCC
is currently covered by the GPL, not GPL plus exception, as described
in PR other/28002.

Janis


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]