[PATCH] MIPS32 DSP intrinsics

[Chao-ying Fu]

  Yes, the CPU's endianness affects which value is stored in a register
for the DSP vector, too.
  Thanks a lot for your help!  I will focus on patching GNU sim for MIPS DSP
instructions.

Regards,
Chao-ying

----- Original Message ----- 
From: "Richard Sandiford" <richard@codesourcery.com>
To: "Chao-ying Fu" <fu@mips.com>
Cc: "Thekkath, Radhika" <radhika@mercury.mips.com>;
<gcc-patches@gcc.gnu.org>; <echristo@apple.com>
Sent: Thursday, July 21, 2005 11:45 PM
Subject: Re: [PATCH] MIPS32 DSP intrinsics


> "Chao-ying Fu" <fu@mips.com> writes:
> > Yes, I could change instruction names in "mips-dsp.md", as you said.
> > Also, I removed "RejectNegative" to have "-mno-dsp".
>
> Thanks Chao-ying.  I also like what you've done with the other bits
> I asked to be changed: the definition of BPOSGE_BUILTIN is much
> clearer now and the changes to mips_expand_direct_builtin are
> much cleaner.
>
> The documentation is a lot better too.  I'll probably make one or two
> minor changes before committing (formatting, etc.), but nothing that
> needs another round of patch/review.
>
> Can I just confirm that this "Note:" in the existing paired-single
> documentation:
>
> -----------------------------------------------------------------------
> @code{v2sf} values are initialized in the same way as aggregates.
> For example:
>
> @smallexample
> v2sf a = @{1.5, 9.1@};
> v2sf b;
> float e, f;
> b = (v2sf) @{e, f@};
> @end smallexample
>
> @emph{Note:} The CPU's endianness determines which value is stored in
> the upper half of a register and which value is stored in the lower half.
> On little-endian targets, the first value is the lower one and the second
> value is the upper one.  The opposite order applies to big-endian targets.
> For example, the code above will set the lower half of @code{a} to
> @code{1.5} on little-endian targets and @code{9.1} on big-endian targets.
> -----------------------------------------------------------------------
>
> applies to the DSP vectors too?  If so, I'll add some sort of note
> before committing.
>
> I'll give Eric (cc'ed) time to comment or object, but this looks good to
> go from my point of view.  Thanks for your contribution, and for your
> patience throughout.
>
> Richard
>