This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH], Optimize (double)vec_extract(n) on PowerPC
- From: Michael Meissner <meissner at linux dot vnet dot ibm dot com>
- To: Segher Boessenkool <segher at kernel dot crashing dot org>
- Cc: Michael Meissner <meissner at linux dot vnet dot ibm dot com>, gcc-patches at gcc dot gnu dot org, David Edelsohn <dje dot gcc at gmail dot com>, Bill Schmidt <wschmidt at linux dot vnet dot ibm dot com>
- Date: Mon, 31 Oct 2016 19:29:09 -0400
- Subject: Re: [PATCH], Optimize (double)vec_extract(n) on PowerPC
- Authentication-results: sourceware.org; auth=none
- References: <20161031212556.GA9310@ibm-tiger.the-meissners.org> <20161031230447.GB3328@gate.crashing.org>
On Mon, Oct 31, 2016 at 06:04:47PM -0500, Segher Boessenkool wrote:
> On Mon, Oct 31, 2016 at 05:25:56PM -0400, Michael Meissner wrote:
> > +(define_insn "vsx_xvcvsxwdp_df"
> > + [(set (match_operand:DF 0 "vsx_register_operand" "=ws")
> > + (unspec:DF [(match_operand:V4SI 1 "vsx_register_operand" "wa")]
> > + UNSPEC_VSX_CVSXWDP))]
> > + "VECTOR_UNIT_VSX_P (V2DFmode)"
> > + "xvcvsxwdp %x0,%x1"
> > + [(set_attr "type" "vecdouble")])
>
> I think the condition here is a pasto? It probably works, but it isn't
> exactly clear what it means here.
I will change it to TARGET_VSX, since it needs the XVCVSXWDP instruction to be
defined.
> > +;; Opimize <type> f = (<type>) vec_extract (vi, <n>)
>
> Typo (optimize).
Thanks.
> > +;; where <type> is a hardware supported binary floating point type
> > +;; that is not double
>
> Full stop.
Ok.
>
> Looks fine otherwise. Please apply with those things taking care of.
>
> Thanks,
>
>
> Segher
>
--
Michael Meissner, IBM
IBM, M/S 2506R, 550 King Street, Littleton, MA 01460-6245, USA
email: meissner@linux.vnet.ibm.com, phone: +1 (978) 899-4797