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 7A1A93858D37; Thu, 31 Aug 2023 11:48:30 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 7A1A93858D37
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1693482510;
	bh=v3sdwjh4M+ijXeInVAovnocRiIzHtNl5KPais5UgoQs=;
	h=From:To:Subject:Date:From;
	b=QRoiaGRV4MFd/wqOidw24pmi6fkpOWrzyTnYrl2EsTUrQYT32k01AtmnB1L1BkDCU
	 6VFEjjR46dLQXpLDN/gaeCT4TB9iD9xRgEFtGrRseO/9p7tOdrph4Pp549PoCC8TtK
	 QDRS2oHFPmjDM0K4R5id15z3o3DhRv2/Pcc44F2w=
From: "kristerw at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/111257] New: new signed overflow after
 vectorizer
Date: Thu, 31 Aug 2023 11:48:29 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: new
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 14.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: kristerw 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: bug_id short_desc product version bug_status
 bug_severity priority component assigned_to reporter target_milestone
Message-ID: <bug-111257-4@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=3D111257

            Bug ID: 111257
           Summary: new signed overflow after vectorizer
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: kristerw at gcc dot gnu.org
  Target Milestone: ---

The vectorizer is not removing the original scalar calculations, and they m=
ay
overflow after vectorization.

This can be seen with

  int a[8];

  void foo(void)
  {
    for (int i =3D 0; i < 8; i++)
      a[i] =3D a[i] + 5;
  }

The IR for the loop before vectorization looks like

  <bb 3> [local count: 954449104]:
  # i_10 =3D PHI <i_7(5), 0(2)>
  # ivtmp_4 =3D PHI <ivtmp_3(5), 8(2)>
  _1 =3D a[i_10];
  _2 =3D _1 + 5;
  a[i_10] =3D _2;
  i_7 =3D i_10 + 1;
  ivtmp_3 =3D ivtmp_4 - 1;
  if (ivtmp_3 !=3D 0)
    goto <bb 5>; [87.50%]
  else
    goto <bb 4>; [12.50%]

  <bb 5> [local count: 835156385]:
  goto <bb 3>; [100.00%]

and it is vectorized to

  <bb 3> [local count: 238585440]:
  # i_10 =3D PHI <i_7(5), 0(2)>
  # ivtmp_4 =3D PHI <ivtmp_3(5), 8(2)>
  # vectp_a.4_9 =3D PHI <vectp_a.4_8(5), &a(2)>
  # vectp_a.8_16 =3D PHI <vectp_a.8_17(5), &a(2)>
  # ivtmp_19 =3D PHI <ivtmp_20(5), 0(2)>
  vect__1.6_13 =3D MEM <vector(4) int> [(int *)vectp_a.4_9];
  _1 =3D a[i_10];
  vect__2.7_15 =3D vect__1.6_13 + { 5, 5, 5, 5 };
  _2 =3D _1 + 5;
  MEM <vector(4) int> [(int *)vectp_a.8_16] =3D vect__2.7_15;
  i_7 =3D i_10 + 1;
  ivtmp_3 =3D ivtmp_4 - 1;
  vectp_a.4_8 =3D vectp_a.4_9 + 16;
  vectp_a.8_17 =3D vectp_a.8_16 + 16;
  ivtmp_20 =3D ivtmp_19 + 1;
  if (ivtmp_20 < 2)
    goto <bb 5>; [50.00%]
  else
    goto <bb 4>; [50.00%]

  <bb 5> [local count: 119292723]:
  goto <bb 3>; [100.00%]

This vectorized loop still read _1 from a[i_10] and adds 5 to it, so the se=
cond
loop iteration will add 5 to the value of a[1]. But the first iteration has
already added 5 to a[1], so we are now doing a different calculation compar=
ed
to the original loop, and this can overflow even if the original did not.=