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 2A4663857C5B; Tue,  5 Jan 2021 09:57:48 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 2A4663857C5B
From: "crazylht at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/98365] Miss vectoization for signed char ifcvt
Date: Tue, 05 Jan 2021 09:57:48 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 11.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: crazylht at gmail dot com
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-98365-4-BdLkzWWI1j@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-98365-4@http.gcc.gnu.org/bugzilla/>
References: <bug-98365-4@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: Tue, 05 Jan 2021 09:57:48 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D98365

--- Comment #4 from Hongtao.liu <crazylht at gmail dot com> ---
> I hope vectorizer reduction can handle the upper sequence.

After hacked in ifcvt, got

.165.cvt
----
  <bb 3> [local count: 1057206201]:
  # cnt_21 =3D PHI <cnt_9(7), 0(15)>
  # i_22 =3D PHI <i_17(7), 0(15)>
  # ivtmp_19 =3D PHI <ivtmp_18(7), 64(15)>
  _1 =3D (sizetype) i_22;
  _2 =3D a_14(D) + _1;
  _3 =3D *_2;
  _5 =3D c_15(D) + _1;
  _6 =3D *_5;
  cnt.1_7 =3D (unsigned char) cnt_21;
  _ifc__35 =3D _3 =3D=3D _6 ? 1 : 0;
  _nop__36 =3D cnt.1_7 + _ifc__35;
  cnt_9 =3D (char) _nop__36;
  i_17 =3D i_22 + 1;
  ivtmp_18 =3D ivtmp_19 - 1;
  if (ivtmp_18 !=3D 0)
    goto <bb 7>; [98.44%]
-------

And successully vectorized.

.166t.vect
------
  <bb 3> [local count: 33071249]:
  # cnt_21 =3D PHI <cnt_9(7), 0(2)>
  # i_22 =3D PHI <i_17(7), 0(2)>
  # ivtmp_19 =3D PHI <ivtmp_18(7), 64(2)>
  # vect_cnt_21.6_38 =3D PHI <vect_cnt_9.16_51(7), { 0, 0, 0, 0, 0, 0, 0, 0=
, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }(2)>
  # vectp_a.7_39 =3D PHI <vectp_a.7_40(7), a_14(D)(2)>
  # vectp_c.10_42 =3D PHI <vectp_c.10_43(7), c_15(D)(2)>
  # ivtmp_56 =3D PHI <ivtmp_57(7), 0(2)>
  _1 =3D (sizetype) i_22;
  _2 =3D a_14(D) + _1;
  vect__3.9_41 =3D MEM <vector(32) char> [(char *)vectp_a.7_39];
  _3 =3D *_2;
  _5 =3D c_15(D) + _1;
  vect__6.12_44 =3D MEM <vector(32) char> [(char *)vectp_c.10_42];
  _6 =3D *_5;
  vect_cnt.13_45 =3D VIEW_CONVERT_EXPR<vector(32) unsigned
char>(vect_cnt_21.6_38);
  cnt.1_7 =3D (unsigned char) cnt_21;
  _48 =3D vect__3.9_41 =3D=3D vect__6.12_44;
  vect__ifc__35.14_49 =3D VEC_COND_EXPR <_48, vect_cst__46, vect_cst__47>;
  _ifc__35 =3D _3 =3D=3D _6 ? 1 : 0;
  vect__nop__36.15_50 =3D vect_cnt.13_45 + vect__ifc__35.14_49;
  _nop__36 =3D cnt.1_7 + _ifc__35;
  vect_cnt_9.16_51 =3D VIEW_CONVERT_EXPR<vector(32) char>(vect__nop__36.15_=
50);
  cnt_9 =3D (char) _nop__36;
  i_17 =3D i_22 + 1;
  ivtmp_18 =3D ivtmp_19 - 1;
  vectp_a.7_40 =3D vectp_a.7_39 + 32;
  vectp_c.10_43 =3D vectp_c.10_42 + 32;
  ivtmp_57 =3D ivtmp_56 + 1;
  if (ivtmp_57 < 2)
    goto <bb 7>; [50.00%]=