From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
 id B1CCD3858032; Wed, 19 Jan 2022 03:30:20 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org B1CCD3858032
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/104049] [12 Regression] vec_select to subreg
 lowering causes superfluous moves
Date: Wed, 19 Jan 2022 03:30:20 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: rtl-optimization
X-Bugzilla-Version: 12.0
X-Bugzilla-Keywords: missed-optimization, ra
X-Bugzilla-Severity: normal
X-Bugzilla-Who: pinskia at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P1
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 12.0
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-104049-4-pXPra7d5Ol@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-104049-4@http.gcc.gnu.org/bugzilla/>
References: <bug-104049-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-BeenThere: gcc-bugs@gcc.gnu.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Gcc-bugs mailing list <gcc-bugs.gcc.gnu.org>
List-Unsubscribe: <https://gcc.gnu.org/mailman/options/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=unsubscribe>
List-Archive: <https://gcc.gnu.org/pipermail/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-request@gcc.gnu.org?subject=help>
List-Subscribe: <https://gcc.gnu.org/mailman/listinfo/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=subscribe>
X-List-Received-Date: Wed, 19 Jan 2022 03:30:20 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D104049

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---

Before my patch we had on the gimple level:

  vect_sum_26.13_38 =3D VIEW_CONVERT_EXPR<vector(4) int>(vect__7.9_48);
  _36 =3D (vector(4) unsigned int) vect_sum_26.13_38;
  _35 =3D .REDUC_PLUS (_36);
  _34 =3D (int) _35;
  _44 =3D .REDUC_PLUS (vect__7.9_48);
  _20 =3D _34 & 65535;
  _10 =3D (unsigned int) _20;
  _12 =3D _44 >> 16;
  _13 =3D _10 + _12;
  _14 =3D _13 >> 1;
  _23 =3D (int) _14;

After we get:

  _43 =3D .REDUC_PLUS (vect__7.11_47);
  _37 =3D _43 & 65535;
  _12 =3D _43 >> 16;
  _13 =3D _12 + _37;
  _14 =3D _13 >> 1;
  _23 =3D (int) _14;

Which is obviously better really. So there is nothing to be done on the gim=
ple
level really. the only thing I can think of is how reducative plus is
implemented.=