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 718FB3858C55; Wed, 21 Sep 2022 00:42:48 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 718FB3858C55
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1663720968;
	bh=2AaQ1SJtnDoj7bqKNpFZo+Ys/I3mcgv1qO5y90x9Iu8=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=W5hJ5XUZW/Ot01WCAEguXrtzsSPkeWRAlugaRly64QQem/VB9AoE9JV3Ft623oiZk
	 YYokUVzjzhy3tfV4doD9RjBBO69N0JE5R/qJ2WdLjTzSnB+cJlSd/6KC0a74Zfgdyj
	 57Ka1lZuZCoIVy4Ra8abmZxC0H+niHX4nvqI3h9A=
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/106989] GCC fail to vectorize and clang
 succeed
Date: Wed, 21 Sep 2022 00:42:48 +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: 13.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: pinskia at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
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-106989-4-2H2AP62gO3@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-106989-4@http.gcc.gnu.org/bugzilla/>
References: <bug-106989-4@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
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D106989

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
/app/example.cpp:20:25: note:   Detected interleaving store a[i_27] and a[i=
_27]
/app/example.cpp:20:25: note:   Queuing group with duplicate access for fix=
up
/app/example.cpp:20:25: note:   zero step in outer loop.
/app/example.cpp:20:25: note:   zero step in outer loop.
/app/example.cpp:20:25: missed:   not vectorized: complicated access patter=
n.
/app/example.cpp:22:18: missed:   not vectorized: complicated access patter=
n.
/app/example.cpp:20:25: missed:  bad data access.
...

/app/example.cpp:21:27: note:   dependence distance  =3D 0.
/app/example.cpp:21:27: note:   dependence distance =3D=3D 0 between b[i_27=
] and
b[i_27]
/app/example.cpp:21:27: note:   dependence distance  =3D 1.
/app/example.cpp:22:18: missed:   not vectorized, possible dependence betwe=
en
data-refs a[i_27] and a[_9]
/app/example.cpp:21:27: missed:  bad data dependence.
/app/example.cpp:21:27: note:  ***** Analysis  failed with vector mode V4SF

There is a missing DSE before hand:
  # VUSE <.MEM_28>
  _1 =3D bD.3768[i_27];
  # VUSE <.MEM_28>
  _2 =3D cD.3769[i_27];
  # VUSE <.MEM_28>
  _3 =3D dD.3770[i_27];
  _4 =3D _2 * _3;
  _5 =3D _1 + _4;
  # .MEM_19 =3D VDEF <.MEM_28>
  aD.3767[i_27] =3D _5;
  # VUSE <.MEM_19>
  _6 =3D eD.3771[i_27];
  _7 =3D _3 * _6;
  _8 =3D _5 + _7;
  # .MEM_20 =3D VDEF <.MEM_19>
  bD.3768[i_27] =3D _8;
  # RANGE [irange] int [1, 31999] NONZERO 0x7fff
  _9 =3D i_27 + 1;
  # VUSE <.MEM_20>
  _10 =3D aD.3767[_9];
  _11 =3D _3 * _10;
  _12 =3D _8 + _11;
  # .MEM_21 =3D VDEF <.MEM_20>
  aD.3767[i_27] =3D _12;

DSE does not notice the store defining MEM_19 does touch the load:
  # VUSE <.MEM_20>
  _10 =3D aD.3767[_9];

And that it is redudent with the store defining MEM_21.=