[Bug c++/103007] New: ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722

[Bug c++/103007] New: ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

            Bug ID: 103007
           Summary: ice in vect_normalize_conj_loc, at
                    tree-vect-slp-patterns.c:722
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

Created attachment 51705
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51705&action=edit
gzipped C++ source code

The attached C++ code, with compiler flag -O2 -march=bdver2, does this:

during GIMPLE pass: slp
Adanaxis/AdanaxisPieceProjectile.cpp: In member function ‘virtual void
AdanaxisPieceProjectile::Move(MushGameLogic&, Mushware::tVal)’:
Adanaxis/AdanaxisPieceProjectile.cpp:123:1: internal compiler error: in
vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722
0x8f64ab vect_normalize_conj_loc
        ../../trunk.git/gcc/tree-vect-slp-patterns.c:722

The bug seems to start sometime between git hash b343a29dbcbfc5ea
and 70c947e4dfaa6d63.

I will have my usual go at reducing the code.

[Bug c++/103007] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #1 from David Binderman <dcb314 at hotmail dot com> ---
Reduced C++ code seems to be:

template <int D> class MushMeshVector {
public:
  MushMeshVector(float, float, float, float);
  float Z() {
    float __trans_tmp_3;
    int inIndex = 2;
    __trans_tmp_3 = m_value[inIndex];
    return __trans_tmp_3;
  }
  float W() {
    float __trans_tmp_4;
    int inIndex = 3;
    __trans_tmp_4 = m_value[inIndex];
    return __trans_tmp_4;
  }
  void ZSet(float inValue) {
    int inIndex = 2;
    m_value[inIndex] = inValue;
  }
  void WSet(float inValue) {
    int inIndex = 3;
    m_value[inIndex] = inValue;
  }
  float m_value[D];
};
template <class> class MushMeshQuaternion : MushMeshVector<4> {
public:
  void PreMultiplyVector(MushMeshVector &) const;
  void PostMultiplyVector(MushMeshVector &) const;
};
template <class T>
void MushMeshQuaternion<T>::PreMultiplyVector(MushMeshVector &ioVec) const {
  ioVec.ZSet(2 - m_value[1] + m_value[3] * 0);
  ioVec.WSet(m_value[3] + m_value[1] - m_value[2] * 0);
}
template <class T>
void MushMeshQuaternion<T>::PostMultiplyVector(MushMeshVector &ioVec) const {
  T c = ioVec.Z(), d = ioVec.W();
  ioVec.ZSet(0 - d * m_value[1]);
  ioVec.WSet(2 - c * m_value[1]);
}
template <class> class MushMeshQuaternionPair {
public:
  void VectorRotate(MushMeshVector<4> &) const;
  MushMeshQuaternion<float> m_first;
  MushMeshQuaternion<float> m_second;
};
template <class T>
void MushMeshQuaternionPair<T>::VectorRotate(MushMeshVector<4> &ioVec) const {
  m_first.PreMultiplyVector(ioVec);
  m_second.PostMultiplyVector(ioVec);
}
class MushMeshPosticity {
public:
  MushMeshQuaternionPair<float> AngPos();
};
class MushGamePiece {
public:
  MushMeshPosticity Post();
  void PostWRef();
};
class AdanaxisPieceProjectile : MushGamePiece {
  void Move();
  float m_acceleration;
};
void AdanaxisPieceProjectile::Move() {
  MushMeshVector<4> accVec(0, 0, 0, m_acceleration);
  Post().AngPos().VectorRotate(accVec);
  PostWRef();
}

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |12.0

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-10-30

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed, reduced to:
typedef float MushMeshVector[4];
struct MushMeshQuaternionPair {
  void VectorRotate(MushMeshVector &);
  MushMeshVector m_first;
  MushMeshVector m_second;
};
void 
MushMeshQuaternionPair::
VectorRotate(MushMeshVector &ioVec)  {
  ioVec[2] = (2 - m_first[1] + m_first[3] * 0);
  ioVec[3] = (m_first[3] + m_first[1] - m_first[2] * 0);
  float c = ioVec[2], d = ioVec[3];
  ioVec[2] = (0 - d * m_second[1]);
  ioVec[3] = (2 - c * m_second[1]);
}

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marxin at gcc dot gnu.org
            Summary|[12 Regression] ice in      |[12 Regression] ice in
                   |vect_normalize_conj_loc, at |vect_normalize_conj_loc, at
                   |tree-vect-slp-patterns.c:72 |tree-vect-slp-patterns.c:72
                   |2                           |2 since
                   |                            |r12-4785-ged3de62ac949c92ad
                   |                            |41ef6de7cc926fbb2a510ce
             Blocks|                            |26163

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
Btw. it affects quite some SPEC benchmarks.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
[Bug 26163] [meta-bug] missed optimization in SPEC (2k17, 2k and 2k6 and 95)

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #4 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Yes a lane check size is missing from fms, will be fixed in a bit.

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |tnfchris at gcc dot gnu.org

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[juergen.reuter at desy dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Jürgen Reuter <juergen.reuter at desy dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |juergen.reuter at desy dot de

--- Comment #5 from Jürgen Reuter <juergen.reuter at desy dot de> ---
I can confirm this ICE also with (legacy) Fortran code, cf. below.
The full code is attached. 

C***********************************************************************

C...PYQQBH
C...Calculates the matrix element for the processes
C...g + g or q + qbar -> Q + Qbar + H (normally with Q = t).
C...REDUCE output and part of the rest courtesy Z. Kunszt, see
C...Z. Kunszt, Nucl. Phys. B247 (1984) 339.

      SUBROUTINE PYQQBH(WTQQBH)

C...Double precision and integer declarations.
      IMPLICIT DOUBLE PRECISION(A-H, O-Z)
      IMPLICIT INTEGER(I-N)
      INTEGER PYK,PYCHGE,PYCOMP
C...Commonblocks.
      COMMON/PYDAT1/MSTU(200),PARU(200),MSTJ(200),PARJ(200)
      COMMON/PYDAT2/KCHG(500,4),PMAS(500,4),PARF(2000),VCKM(4,4)
      COMMON/PYPARS/MSTP(200),PARP(200),MSTI(200),PARI(200)
      COMMON/PYINT1/MINT(400),VINT(400)
      COMMON/PYINT2/ISET(500),KFPR(500,2),COEF(500,20),ICOL(40,4,2)
      SAVE /PYDAT1/,/PYDAT2/,/PYPARS/,/PYINT1/,/PYINT2/
C...Local arrays and function.
      DIMENSION PP(15,4),CLR(8,8),FM(10,10),RM(8,8),DX(8)
      DOT(I,J)=PP(I,4)*PP(J,4)-PP(I,1)*PP(J,1)-PP(I,2)*PP(J,2)-
     &PP(I,3)*PP(J,3)

C...Mass parameters.
      WTQQBH=0D0
      ISUB=MINT(1)
      SHPR=SQRT(VINT(26))*VINT(1)
      PQ=PMAS(PYCOMP(KFPR(ISUB,2)),1)
      PH=SQRT(VINT(21))*VINT(1)
      SPQ=PQ**2
      SPH=PH**2

C...Set up outgoing kinematics: 1=t, 2=tbar, 3=H.
      DO 100 I=1,2
        PT=SQRT(MAX(0D0,VINT(197+5*I)))
        PP(I,1)=PT*COS(VINT(198+5*I))
        PP(I,2)=PT*SIN(VINT(198+5*I))
  100 CONTINUE
      PP(3,1)=-PP(1,1)-PP(2,1)
      PP(3,2)=-PP(1,2)-PP(2,2)
      PMS1=SPQ+PP(1,1)**2+PP(1,2)**2
      PMS2=SPQ+PP(2,1)**2+PP(2,2)**2
      PMS3=SPH+PP(3,1)**2+PP(3,2)**2
      PMT3=SQRT(PMS3)
      PP(3,3)=PMT3*SINH(VINT(211))
      PP(3,4)=PMT3*COSH(VINT(211))
      PMS12=(SHPR-PP(3,4))**2-PP(3,3)**2
      PP(1,3)=(-PP(3,3)*(PMS12+PMS1-PMS2)+
     &VINT(213)*(SHPR-PP(3,4))*VINT(220))/(2D0*PMS12)
      PP(2,3)=-PP(1,3)-PP(3,3)
      PP(1,4)=SQRT(PMS1+PP(1,3)**2)
      PP(2,4)=SQRT(PMS2+PP(2,3)**2)

C...Set up incoming kinematics and derived momentum combinations.
      DO 110 I=4,5
        PP(I,1)=0D0
        PP(I,2)=0D0
        PP(I,3)=-0.5D0*SHPR*(-1)**I
        PP(I,4)=-0.5D0*SHPR
  110 CONTINUE
      DO 120 J=1,4
        PP(6,J)=PP(1,J)+PP(2,J)
        PP(7,J)=PP(1,J)+PP(3,J)
        PP(8,J)=PP(1,J)+PP(4,J)
        PP(9,J)=PP(1,J)+PP(5,J)
        PP(10,J)=-PP(2,J)-PP(3,J)
        PP(11,J)=-PP(2,J)-PP(4,J)
        PP(12,J)=-PP(2,J)-PP(5,J)
        PP(13,J)=-PP(4,J)-PP(5,J)
  120 CONTINUE

C...Derived kinematics invariants.
      X1=DOT(1,2)
      X2=DOT(1,3)
      X3=DOT(1,4)
      X4=DOT(1,5)
      X5=DOT(2,3)
      X6=DOT(2,4)
      X7=DOT(2,5)
      X8=DOT(3,4)
      X9=DOT(3,5)
      X10=DOT(4,5)

C...Propagators.
      SS1=DOT(7,7)-SPQ
      SS2=DOT(8,8)-SPQ
      SS3=DOT(9,9)-SPQ
      SS4=DOT(10,10)-SPQ
      SS5=DOT(11,11)-SPQ
      SS6=DOT(12,12)-SPQ
      SS7=DOT(13,13)
      DX(1)=SS1*SS6
      DX(2)=SS2*SS6
      DX(3)=SS2*SS4
      DX(4)=SS1*SS5
      DX(5)=SS3*SS5
      DX(6)=SS3*SS4
      DX(7)=SS7*SS1
      DX(8)=SS7*SS4

C...Define colour coefficients for g + g -> Q + Qbar + H.
      IF(ISUB.EQ.121.OR.ISUB.EQ.181.OR.ISUB.EQ.186) THEN
        DO 140 I=1,3
          DO 130 J=1,3
            CLR(I,J)=16D0/3D0
            CLR(I+3,J+3)=16D0/3D0
            CLR(I,J+3)=-2D0/3D0
            CLR(I+3,J)=-2D0/3D0
  130     CONTINUE
  140   CONTINUE
        DO 160 L=1,2
          DO 150 I=1,3
            CLR(I,6+L)=-6D0
            CLR(I+3,6+L)=6D0
            CLR(6+L,I)=-6D0
            CLR(6+L,I+3)=6D0
  150     CONTINUE
  160   CONTINUE
        DO 180 K1=1,2
          DO 170 K2=1,2
            CLR(6+K1,6+K2)=12D0
  170     CONTINUE
  180   CONTINUE

C...Evaluate matrix elements for g + g -> Q + Qbar + H.
        FM(1,1)=64*PQ**6+16*PQ**4*PH**2+32*PQ**4*(X1+2*X2+X4+X9+2*
     &  X7+X5)+8*PQ**2*PH**2*(-X1-X4+2*X7)+16*PQ**2*(X2*X9+4*X2*
     &  X7+X2*X5-2*X4*X7-2*X9*X7)+8*PH**2*X4*X7-16*X2*X9*X7
        FM(1,2)=16*PQ**6+8*PQ**4*(-2*X1+X2-2*X3-2*X4-4*X10+X9-X8+2
     &  *X7-4*X6+X5)+8*PQ**2*(-2*X1*X2-2*X2*X4-2*X2*X10+X2*X7-2*
     &  X2*X6-2*X3*X7+2*X4*X7+4*X10*X7-X9*X7-X8*X7)+16*X2*X7*(X4+
     &  X10)
        FM(1,3)=16*PQ**6-4*PQ**4*PH**2+8*PQ**4*(-2*X1+2*X2-2*X3-4*
     &  X4-8*X10+X9+X8-2*X7-4*X6+2*X5)-(4*PQ**2*PH**2)*(X1+X4+X10
     &  +X6)+8*PQ**2*(-2*X1*X2-2*X1*X10+X1*X9+X1*X8-2*X1*X5+X2**2
     &  -4*X2*X4-5*X2*X10+X2*X8-X2*X7-3*X2*X6+X2*X5+X3*X9+2*X3*X7
     &  -X3*X5+X4*X8+2*X4*X6-3*X4*X5-5*X10*X5+X9*X8+X9*X6+X9*X5+
     &  X8*X7-4*X6*X5+X5**2)-(16*X2*X5)*(X1+X4+X10+X6)
        FM(1,4)=16*PQ**6+4*PQ**4*PH**2+16*PQ**4*(-X1+X2-X3-X4+X10-
     &  X9-X8+2*X7+2*X6-X5)+4*PQ**2*PH**2*(X1+X3+X4+X10+2*X7+2*X6
     &  )+8*PQ**2*(4*X1*X10+4*X1*X7+4*X1*X6+2*X2*X10-X2*X9-X2*X8+
     &  4*X2*X7+4*X2*X6-X2*X5+4*X10*X5+4*X7*X5+4*X6*X5)-(8*PH**2*
     &  X1)*(X10+X7+X6)+16*X2*X5*(X10+X7+X6)
        FM(1,5)=8*PQ**4*(-2*X1-2*X4+X10-X9)+4*PQ**2*(4*X1**2-2*X1*
     &  X2+8*X1*X3+6*X1*X10-2*X1*X9+4*X1*X8+4*X1*X7+4*X1*X6+2*X1*
     &  X5+X2*X10+4*X3*X4-X3*X9+2*X3*X7+3*X4*X8-2*X4*X6+2*X4*X5-4
     &  *X10*X7+3*X10*X5-3*X9*X6+3*X8*X7-4*X7**2+4*X7*X5)+8*(X1**
     &  2*X9-X1**2*X8-X1*X2*X7+X1*X2*X6+X1*X3*X9+X1*X3*X5-X1*X4*
     &  X8-X1*X4*X5+X1*X10*X9+X1*X9*X7+X1*X9*X6-X1*X8*X7-X2*X3*X7
     &  +X2*X4*X6-X2*X10*X7-X2*X7**2+X3*X7*X5-X4*X10*X5-X4*X7*X5-
     &  X4*X6*X5)
        FM(1,6)=16*PQ**4*(-4*X1-X4+X9-X7)+4*PQ**2*PH**2*(-2*X1-X4-
     &  X7)+16*PQ**2*(-2*X1**2-3*X1*X2-2*X1*X4-3*X1*X9-2*X1*X7-3*
     &  X1*X5-2*X2*X4-2*X7*X5)-8*PH**2*X4*X7+8*(-X1*X2*X9-2*X1*X2
     &  *X5-X1*X9**2-X1*X9*X5+X2**2*X7-X2*X4*X5+X2*X9*X7-X2*X7*X5
     &  +X4*X9*X5+X4*X5**2)
        FM(1,7)=8*PQ**4*(2*X3+X4+3*X10+X9+2*X8+3*X7+6*X6)+2*PQ**2*
     &  PH**2*(-2*X3-X4+3*X10+3*X7+6*X6)+4*PQ**2*(4*X1*X10+4*X1*
     &  X7+8*X1*X6+6*X2*X10+X2*X9+2*X2*X8+6*X2*X7+12*X2*X6-8*X3*
     &  X7+4*X4*X7+4*X4*X6+4*X10*X5+4*X9*X7+4*X9*X6-8*X8*X7+4*X7*
     &  X5+8*X6*X5)+4*PH**2*(-X1*X10-X1*X7-2*X1*X6+2*X3*X7-X4*X7-
     &  X4*X6)+8*X2*(X10*X5+X9*X7+X9*X6-2*X8*X7+X7*X5+2*X6*X5)
        FM(1,8)=8*PQ**4*(2*X3+X4+3*X10+2*X9+X8+3*X7+6*X6)+2*PQ**2*
     &  PH**2*(-2*X3-X4+2*X10+X7+2*X6)+4*PQ**2*(4*X1*X10-2*X1*X9+
     &  2*X1*X8+4*X1*X7+8*X1*X6+5*X2*X10+2*X2*X9+X2*X8+4*X2*X7+8*
     &  X2*X6-X3*X9-8*X3*X7+2*X3*X5+2*X4*X9-X4*X8+4*X4*X7+4*X4*X6
     &  +4*X4*X5+5*X10*X5+X9**2-X9*X8+2*X9*X7+5*X9*X6+X9*X5-7*X8*
     &  X7+2*X8*X5+2*X7*X5+10*X6*X5)+2*PH**2*(-X1*X10+X3*X7-2*X4*
     &  X7+X4*X6)+4*(-X1*X9**2+X1*X9*X8-2*X1*X9*X5-X1*X8*X5+2*X2*
     &  X10*X5+X2*X9*X7+X2*X9*X6-2*X2*X8*X7+3*X2*X6*X5+X3*X9*X5+
     &  X3*X5**2+X4*X9*X5-2*X4*X8*X5+2*X4*X5**2)
        FM(2,2)=16*PQ**6+16*PQ**4*(-X1+X3-X4-X10+X7-X6)+16*PQ**2*(
     &  X3*X10+X3*X7+X3*X6+X4*X7+X10*X7)-16*X3*X10*X7
        FM(2,3)=16*PQ**6+8*PQ**4*(-2*X1+X2+2*X3-4*X4-4*X10-X9+X8-2
     &  *X7-2*X6+X5)+8*PQ**2*(-2*X1*X5+4*X3*X10-X3*X9-X3*X8-2*X3*
     &  X7+2*X3*X6+X3*X5-2*X4*X5-2*X10*X5-2*X6*X5)+16*X3*X5*(X10+
     &  X6)
        FM(2,4)=8*PQ**4*(-2*X1-2*X3+X10-X8)+4*PQ**2*(4*X1**2-2*X1*
     &  X2+8*X1*X4+6*X1*X10+4*X1*X9-2*X1*X8+4*X1*X7+4*X1*X6+2*X1*
     &  X5+X2*X10+4*X3*X4+3*X3*X9-2*X3*X7+2*X3*X5-X4*X8+2*X4*X6-4
     &  *X10*X6+3*X10*X5+3*X9*X6-3*X8*X7-4*X6**2+4*X6*X5)+8*(-X1
     &  **2*X9+X1**2*X8+X1*X2*X7-X1*X2*X6-X1*X3*X9-X1*X3*X5+X1*X4
     &  *X8+X1*X4*X5+X1*X10*X8-X1*X9*X6+X1*X8*X7+X1*X8*X6+X2*X3*
     &  X7-X2*X4*X6-X2*X10*X6-X2*X6**2-X3*X10*X5-X3*X7*X5-X3*X6*
     &  X5+X4*X6*X5)
        FM(2,5)=16*PQ**4*X10+8*PQ**2*(2*X1**2+2*X1*X3+2*X1*X4+2*X1
     &  *X10+2*X1*X7+2*X1*X6+X3*X7+X4*X6)+8*(-2*X1**3-2*X1**2*X3-
     &  2*X1**2*X4-2*X1**2*X10-2*X1**2*X7-2*X1**2*X6-2*X1*X3*X4-
     &  X1*X3*X10-2*X1*X3*X6-X1*X4*X10-2*X1*X4*X7-X1*X10**2-X1*
     &  X10*X7-X1*X10*X6-2*X1*X7*X6+X3**2*X7-X3*X4*X7-X3*X4*X6+X3
     &  *X10*X7+X3*X7**2-X3*X7*X6+X4**2*X6+X4*X10*X6-X4*X7*X6+X4*
     &  X6**2)
        FM(2,6)=8*PQ**4*(-2*X1+X10-X9-2*X7)+4*PQ**2*(4*X1**2+2*X1*
     &  X2+4*X1*X3+4*X1*X4+6*X1*X10-2*X1*X9+4*X1*X8+8*X1*X6-2*X1*
     &  X5+4*X2*X4+3*X2*X10+2*X2*X7-3*X3*X9-2*X3*X7-4*X4**2-4*X4*
     &  X10+3*X4*X8+2*X4*X6+X10*X5-X9*X6+3*X8*X7+4*X7*X6)+8*(X1**
     &  2*X9-X1**2*X8-X1*X2*X7+X1*X2*X6+X1*X3*X9+X1*X3*X5+X1*X4*
     &  X9-X1*X4*X8-X1*X4*X5+X1*X10*X9+X1*X9*X6-X1*X8*X7-X2*X3*X7
     &  -X2*X4*X7+X2*X4*X6-X2*X10*X7+X3*X7*X5-X4**2*X5-X4*X10*X5-
     &  X4*X6*X5)
        FM(2,7)=8*PQ**4*(X3+2*X4+3*X10+X7+2*X6)+4*PQ**2*(-4*X1*X3-
     &  2*X1*X4-2*X1*X10+X1*X9-X1*X8-4*X1*X7-2*X1*X6+X2*X3+2*X2*
     &  X4+3*X2*X10+X2*X7+2*X2*X6-6*X3*X4-6*X3*X10-2*X3*X9-2*X3*
     &  X7-4*X3*X6-X3*X5-6*X4**2-6*X4*X10-3*X4*X9-X4*X8-4*X4*X7-2
     &  *X4*X6-2*X4*X5-3*X10*X9-3*X10*X8-6*X10*X7-6*X10*X6+X10*X5
     &  +X9*X7-2*X8*X7-2*X8*X6-6*X7*X6+X7*X5-6*X6**2+2*X6*X5)+4*(
     &  -X1**2*X9+X1**2*X8-2*X1*X2*X10-3*X1*X2*X7-3*X1*X2*X6+X1*
     &  X3*X9-X1*X3*X5+X1*X4*X9+X1*X4*X8+X1*X4*X5+X1*X10*X9+X1*
     &  X10*X8-X1*X9*X6+X1*X8*X6+X2*X3*X7-3*X2*X4*X7-X2*X4*X6-3*
     &  X2*X10*X7-3*X2*X10*X6-3*X2*X7*X6-3*X2*X6**2-2*X3*X4*X5-X3
     &  *X10*X5-X3*X6*X5-X4**2*X5-X4*X10*X5+X4*X6*X5)
        FM(2,8)=8*PQ**4*(X3+2*X4+3*X10+X7+2*X6)+4*PQ**2*(-4*X1*X3-
     &  2*X1*X4-2*X1*X10-X1*X9+X1*X8-4*X1*X7-2*X1*X6+X2*X3+2*X2*
     &  X4+X2*X10-X2*X7-2*X2*X6-6*X3*X4-6*X3*X10-2*X3*X9+X3*X8-2*
     &  X3*X7-4*X3*X6+X3*X5-6*X4**2-6*X4*X10-2*X4*X9-4*X4*X7-2*X4
     &  *X6+2*X4*X5-3*X10*X9-3*X10*X8-6*X10*X7-6*X10*X6+3*X10*X5-
     &  X9*X6-2*X8*X7-3*X8*X6-6*X7*X6+X7*X5-6*X6**2+2*X6*X5)+4*(
     &  X1**2*X9-X1**2*X8-X1*X2*X7+X1*X2*X6-3*X1*X3*X5+X1*X4*X9-
     &  X1*X4*X8-3*X1*X4*X5+X1*X10*X9+X1*X10*X8-2*X1*X10*X5+X1*X9
     &  *X6+X1*X8*X7+X1*X8*X6-X2*X4*X7+X2*X4*X6-X2*X10*X7-X2*X10*
     &  X6-2*X2*X7*X6-X2*X6**2-3*X3*X4*X5-3*X3*X10*X5+X3*X7*X5-3*
     &  X3*X6*X5-3*X4**2*X5-3*X4*X10*X5-X4*X6*X5)
        FM(3,3)=64*PQ**6+16*PQ**4*PH**2+32*PQ**4*(X1+X2+2*X3+X8+X6
     &  +2*X5)+8*PQ**2*PH**2*(-X1+2*X3-X6)+16*PQ**2*(X2*X5-2*X3*
     &  X8-2*X3*X6+4*X3*X5+X8*X5)+8*PH**2*X3*X6-16*X3*X8*X5
        FM(3,4)=16*PQ**4*(-4*X1-X3+X8-X6)+4*PQ**2*PH**2*(-2*X1-X3-
     &  X6)+16*PQ**2*(-2*X1**2-3*X1*X2-2*X1*X3-3*X1*X8-2*X1*X6-3*
     &  X1*X5-2*X2*X3-2*X6*X5)-8*PH**2*X3*X6+8*(-X1*X2*X8-2*X1*X2
     &  *X5-X1*X8**2-X1*X8*X5+X2**2*X6-X2*X3*X5+X2*X8*X6-X2*X6*X5
     &  +X3*X8*X5+X3*X5**2)
        FM(3,5)=8*PQ**4*(-2*X1+X10-X8-2*X6)+4*PQ**2*(4*X1**2+2*X1*
     &  X2+4*X1*X3+4*X1*X4+6*X1*X10+4*X1*X9-2*X1*X8+8*X1*X7-2*X1*
     &  X5+4*X2*X3+3*X2*X10+2*X2*X6-4*X3**2-4*X3*X10+3*X3*X9+2*X3
     &  *X7-3*X4*X8-2*X4*X6+X10*X5+3*X9*X6-X8*X7+4*X7*X6)+8*(-X1
     &  **2*X9+X1**2*X8+X1*X2*X7-X1*X2*X6-X1*X3*X9+X1*X3*X8-X1*X3
     &  *X5+X1*X4*X8+X1*X4*X5+X1*X10*X8-X1*X9*X6+X1*X8*X7+X2*X3*
     &  X7-X2*X3*X6-X2*X4*X6-X2*X10*X6-X3**2*X5-X3*X10*X5-X3*X7*
     &  X5+X4*X6*X5)
        FM(3,6)=16*PQ**6+4*PQ**4*PH**2+16*PQ**4*(-X1-X2+2*X3+2*X4+
     &  X10-X9-X8-X7-X6+X5)+4*PQ**2*PH**2*(X1+2*X3+2*X4+X10+X7+X6
     &  )+8*PQ**2*(4*X1*X3+4*X1*X4+4*X1*X10+4*X2*X3+4*X2*X4+4*X2*
     &  X10-X2*X5+4*X3*X5+4*X4*X5+2*X10*X5-X9*X5-X8*X5)-(8*PH**2*
     &  X1)*(X3+X4+X10)+16*X2*X5*(X3+X4+X10)
        FM(3,7)=8*PQ**4*(3*X3+6*X4+3*X10+X9+2*X8+2*X7+X6)+2*PQ**2*
     &  PH**2*(X3+2*X4+2*X10-2*X7-X6)+4*PQ**2*(4*X1*X3+8*X1*X4+4*
     &  X1*X10+2*X1*X9-2*X1*X8+2*X2*X3+10*X2*X4+5*X2*X10+2*X2*X9+
     &  X2*X8+2*X2*X7+4*X2*X6-7*X3*X9+2*X3*X8-8*X3*X7+4*X3*X6+4*
     &  X3*X5+5*X4*X8+4*X4*X6+8*X4*X5+5*X10*X5-X9*X8-X9*X6+X9*X5+
     &  X8**2-X8*X7+2*X8*X6+2*X8*X5)+2*PH**2*(-X1*X10+X3*X7-2*X3*
     &  X6+X4*X6)+4*(-X1*X2*X9-2*X1*X2*X8+X1*X9*X8-X1*X8**2+X2**2
     &  *X7+2*X2**2*X6+3*X2*X4*X5+2*X2*X10*X5-2*X2*X9*X6+X2*X8*X7
     &  +X2*X8*X6-2*X3*X9*X5+X3*X8*X5+X4*X8*X5)
        FM(3,8)=8*PQ**4*(3*X3+6*X4+3*X10+2*X9+X8+2*X7+X6)+2*PQ**2*
     &  PH**2*(3*X3+6*X4+3*X10-2*X7-X6)+4*PQ**2*(4*X1*X3+8*X1*X4+
     &  4*X1*X10+4*X2*X3+8*X2*X4+4*X2*X10-8*X3*X9+4*X3*X8-8*X3*X7
     &  +4*X3*X6+6*X3*X5+4*X4*X8+4*X4*X6+12*X4*X5+6*X10*X5+2*X9*
     &  X5+X8*X5)+4*PH**2*(-X1*X3-2*X1*X4-X1*X10+2*X3*X7-X3*X6-X4
     &  *X6)+8*X5*(X2*X3+2*X2*X4+X2*X10-2*X3*X9+X3*X8+X4*X8)
        FM(4,4)=64*PQ**6+16*PQ**4*PH**2+32*PQ**4*(X1+2*X2+X3+X8+2*
     &  X6+X5)+8*PQ**2*PH**2*(-X1-X3+2*X6)+16*PQ**2*(X2*X8+4*X2*
     &  X6+X2*X5-2*X3*X6-2*X8*X6)+8*PH**2*X3*X6-16*X2*X8*X6
        FM(4,5)=16*PQ**6+8*PQ**4*(-2*X1+X2-2*X3-2*X4-4*X10-X9+X8-4
     &  *X7+2*X6+X5)+8*PQ**2*(-2*X1*X2-2*X2*X3-2*X2*X10-2*X2*X7+
     &  X2*X6+2*X3*X6-2*X4*X6+4*X10*X6-X9*X6-X8*X6)+16*X2*X6*(X3+
     &  X10)
        FM(4,6)=16*PQ**6-4*PQ**4*PH**2+8*PQ**4*(-2*X1+2*X2-4*X3-2*
     &  X4-8*X10+X9+X8-4*X7-2*X6+2*X5)-(4*PQ**2*PH**2)*(X1+X3+X10
     &  +X7)+8*PQ**2*(-2*X1*X2-2*X1*X10+X1*X9+X1*X8-2*X1*X5+X2**2
     &  -4*X2*X3-5*X2*X10+X2*X9-3*X2*X7-X2*X6+X2*X5+X3*X9+2*X3*X7
     &  -3*X3*X5+X4*X8+2*X4*X6-X4*X5-5*X10*X5+X9*X8+X9*X6+X8*X7+
     &  X8*X5-4*X7*X5+X5**2)-(16*X2*X5)*(X1+X3+X10+X7)
        FM(4,7)=8*PQ**4*(-X3-2*X4-3*X10-2*X9-X8-6*X7-3*X6)+2*PQ**2
     &  *PH**2*(X3+2*X4-3*X10-6*X7-3*X6)+4*PQ**2*(-4*X1*X10-8*X1*
     &  X7-4*X1*X6-6*X2*X10-2*X2*X9-X2*X8-12*X2*X7-6*X2*X6-4*X3*
     &  X7-4*X3*X6+8*X4*X6-4*X10*X5+8*X9*X6-4*X8*X7-4*X8*X6-8*X7*
     &  X5-4*X6*X5)+4*PH**2*(X1*X10+2*X1*X7+X1*X6+X3*X7+X3*X6-2*
     &  X4*X6)+8*X2*(-X10*X5+2*X9*X6-X8*X7-X8*X6-2*X7*X5-X6*X5)
        FM(4,8)=8*PQ**4*(-X3-2*X4-3*X10-X9-2*X8-6*X7-3*X6)+2*PQ**2
     &  *PH**2*(X3+2*X4-2*X10-2*X7-X6)+4*PQ**2*(-4*X1*X10-2*X1*X9
     &  +2*X1*X8-8*X1*X7-4*X1*X6-5*X2*X10-X2*X9-2*X2*X8-8*X2*X7-4
     &  *X2*X6+X3*X9-2*X3*X8-4*X3*X7-4*X3*X6-4*X3*X5+X4*X8+8*X4*
     &  X6-2*X4*X5-5*X10*X5+X9*X8+7*X9*X6-2*X9*X5-X8**2-5*X8*X7-2
     &  *X8*X6-X8*X5-10*X7*X5-2*X6*X5)+2*PH**2*(X1*X10-X3*X7+2*X3
     &  *X6-X4*X6)+4*(-X1*X9*X8+X1*X9*X5+X1*X8**2+2*X1*X8*X5-2*X2
     &  *X10*X5+2*X2*X9*X6-X2*X8*X7-X2*X8*X6-3*X2*X7*X5+2*X3*X9*
     &  X5-X3*X8*X5-2*X3*X5**2-X4*X8*X5-X4*X5**2)
        FM(5,5)=16*PQ**6+16*PQ**4*(-X1-X3+X4-X10-X7+X6)+16*PQ**2*(
     &  X3*X6+X4*X10+X4*X7+X4*X6+X10*X6)-16*X4*X10*X6
        FM(5,6)=16*PQ**6+8*PQ**4*(-2*X1+X2-4*X3+2*X4-4*X10+X9-X8-2
     &  *X7-2*X6+X5)+8*PQ**2*(-2*X1*X5-2*X3*X5+4*X4*X10-X4*X9-X4*
     &  X8+2*X4*X7-2*X4*X6+X4*X5-2*X10*X5-2*X7*X5)+16*X4*X5*(X10+
     &  X7)
        FM(5,7)=8*PQ**4*(-2*X3-X4-3*X10-2*X7-X6)+4*PQ**2*(2*X1*X3+
     &  4*X1*X4+2*X1*X10+X1*X9-X1*X8+2*X1*X7+4*X1*X6-2*X2*X3-X2*
     &  X4-3*X2*X10-2*X2*X7-X2*X6+6*X3**2+6*X3*X4+6*X3*X10+X3*X9+
     &  3*X3*X8+2*X3*X7+4*X3*X6+2*X3*X5+6*X4*X10+2*X4*X8+4*X4*X7+
     &  2*X4*X6+X4*X5+3*X10*X9+3*X10*X8+6*X10*X7+6*X10*X6-X10*X5+
     &  2*X9*X7+2*X9*X6-X8*X6+6*X7**2+6*X7*X6-2*X7*X5-X6*X5)+4*(-
     &  X1**2*X9+X1**2*X8+2*X1*X2*X10+3*X1*X2*X7+3*X1*X2*X6-X1*X3
     &  *X9-X1*X3*X8-X1*X3*X5-X1*X4*X8+X1*X4*X5-X1*X10*X9-X1*X10*
     &  X8-X1*X9*X7+X1*X8*X7+X2*X3*X7+3*X2*X3*X6-X2*X4*X6+3*X2*
     &  X10*X7+3*X2*X10*X6+3*X2*X7**2+3*X2*X7*X6+X3**2*X5+2*X3*X4
     &  *X5+X3*X10*X5-X3*X7*X5+X4*X10*X5+X4*X7*X5)
        FM(5,8)=8*PQ**4*(-2*X3-X4-3*X10-2*X7-X6)+4*PQ**2*(2*X1*X3+
     &  4*X1*X4+2*X1*X10-X1*X9+X1*X8+2*X1*X7+4*X1*X6-2*X2*X3-X2*
     &  X4-X2*X10+2*X2*X7+X2*X6+6*X3**2+6*X3*X4+6*X3*X10+2*X3*X8+
     &  2*X3*X7+4*X3*X6-2*X3*X5+6*X4*X10-X4*X9+2*X4*X8+4*X4*X7+2*
     &  X4*X6-X4*X5+3*X10*X9+3*X10*X8+6*X10*X7+6*X10*X6-3*X10*X5+
     &  3*X9*X7+2*X9*X6+X8*X7+6*X7**2+6*X7*X6-2*X7*X5-X6*X5)+4*(
     &  X1**2*X9-X1**2*X8-X1*X2*X7+X1*X2*X6+X1*X3*X9-X1*X3*X8+3*
     &  X1*X3*X5+3*X1*X4*X5-X1*X10*X9-X1*X10*X8+2*X1*X10*X5-X1*X9
     &  *X7-X1*X9*X6-X1*X8*X7-X2*X3*X7+X2*X3*X6+X2*X10*X7+X2*X10*
     &  X6+X2*X7**2+2*X2*X7*X6+3*X3**2*X5+3*X3*X4*X5+3*X3*X10*X5+
     &  X3*X7*X5+3*X4*X10*X5+3*X4*X7*X5-X4*X6*X5)
        FM(6,6)=64*PQ**6+16*PQ**4*PH**2+32*PQ**4*(X1+X2+2*X4+X9+X7
     &  +2*X5)+8*PQ**2*PH**2*(-X1+2*X4-X7)+16*PQ**2*(X2*X5-2*X4*
     &  X9-2*X4*X7+4*X4*X5+X9*X5)+8*PH**2*X4*X7-16*X4*X9*X5
        FM(6,7)=8*PQ**4*(-6*X3-3*X4-3*X10-2*X9-X8-X7-2*X6)+2*PQ**2
     &  *PH**2*(-2*X3-X4-2*X10+X7+2*X6)+4*PQ**2*(-8*X1*X3-4*X1*X4
     &  -4*X1*X10+2*X1*X9-2*X1*X8-10*X2*X3-2*X2*X4-5*X2*X10-X2*X9
     &  -2*X2*X8-4*X2*X7-2*X2*X6-5*X3*X9-4*X3*X7-8*X3*X5-2*X4*X9+
     &  7*X4*X8-4*X4*X7+8*X4*X6-4*X4*X5-5*X10*X5-X9**2+X9*X8-2*X9
     &  *X7+X9*X6-2*X9*X5+X8*X7-X8*X5)+2*PH**2*(X1*X10-X3*X7+2*X4
     &  *X7-X4*X6)+4*(2*X1*X2*X9+X1*X2*X8+X1*X9**2-X1*X9*X8-2*X2
     &  **2*X7-X2**2*X6-3*X2*X3*X5-2*X2*X10*X5-X2*X9*X7-X2*X9*X6+
     &  2*X2*X8*X7-X3*X9*X5-X4*X9*X5+2*X4*X8*X5)
        FM(6,8)=8*PQ**4*(-6*X3-3*X4-3*X10-X9-2*X8-X7-2*X6)+2*PQ**2
     &  *PH**2*(-6*X3-3*X4-3*X10+X7+2*X6)+4*PQ**2*(-8*X1*X3-4*X1*
     &  X4-4*X1*X10-8*X2*X3-4*X2*X4-4*X2*X10-4*X3*X9-4*X3*X7-12*
     &  X3*X5-4*X4*X9+8*X4*X8-4*X4*X7+8*X4*X6-6*X4*X5-6*X10*X5-X9
     &  *X5-2*X8*X5)+4*PH**2*(2*X1*X3+X1*X4+X1*X10+X3*X7+X4*X7-2*
     &  X4*X6)+8*X5*(-2*X2*X3-X2*X4-X2*X10-X3*X9-X4*X9+2*X4*X8)
        FM(7,7)=72*PQ**4*X10+18*PQ**2*PH**2*X10+8*PQ**2*(X1*X10+9*
     &  X2*X10+7*X3*X7+2*X3*X6+2*X4*X7+7*X4*X6+X10*X5+2*X9*X7+7*
     &  X9*X6+7*X8*X7+2*X8*X6)+2*PH**2*(-X1*X10-7*X3*X7-2*X3*X6-2
     &  *X4*X7-7*X4*X6)+4*X2*(X10*X5+2*X9*X7+7*X9*X6+7*X8*X7+2*X8
     &  *X6)
        FM(7,8)=72*PQ**4*X10+2*PQ**2*PH**2*X10+4*PQ**2*(2*X1*X10+
     &  10*X2*X10+7*X3*X9+2*X3*X8+14*X3*X7+4*X3*X6+2*X4*X9+7*X4*
     &  X8+4*X4*X7+14*X4*X6+10*X10*X5+X9**2+7*X9*X8+2*X9*X7+7*X9*
     &  X6+X8**2+7*X8*X7+2*X8*X6)+2*PH**2*(7*X1*X10-7*X3*X7-2*X3*
     &  X6-2*X4*X7-7*X4*X6)+2*(-2*X1*X9**2-14*X1*X9*X8-2*X1*X8**2
     &  +2*X2*X10*X5+2*X2*X9*X7+7*X2*X9*X6+7*X2*X8*X7+2*X2*X8*X6+
     &  7*X3*X9*X5+2*X3*X8*X5+2*X4*X9*X5+7*X4*X8*X5)
        FM(8,8)=72*PQ**4*X10+18*PQ**2*PH**2*X10+8*PQ**2*(X1*X10+X2
     &  *X10+7*X3*X9+2*X3*X8+7*X3*X7+2*X3*X6+2*X4*X9+7*X4*X8+2*X4
     &  *X7+7*X4*X6+9*X10*X5)+2*PH**2*(-X1*X10-7*X3*X7-2*X3*X6-2*
     &  X4*X7-7*X4*X6)+4*X5*(X2*X10+7*X3*X9+2*X3*X8+2*X4*X9+7*X4*
     &  X8)
        FM(9,9)=-4*PQ**4*X10-PQ**2*PH**2*X10+4*PQ**2*(-X1*X10-X2*X10+
     &  X3*X7+X4*X6-X10*X5+X9*X6+X8*X7)+PH**2*(X1*X10-X3*X7-X4*X6
     &  )+2*X2*(-X10*X5+X9*X6+X8*X7)
        FM(9,10)=-4*PQ**4*X10-PQ**2*PH**2*X10+2*PQ**2*(-2*X1*X10-2*X2*
     &  X10+2*X3*X9+2*X3*X7+2*X4*X6-2*X10*X5+X9*X8+2*X8*X7)+PH**2
     &  *(X1*X10-X3*X7-X4*X6)+2*(-X1*X9*X8-X2*X10*X5+X2*X8*X7+X3*
     &  X9*X5)
        FMXX=-4*PQ**4*X10-PQ**2*PH**2*X10+2*PQ**2*(-2*X1*X10-2*X2*
     &  X10+2*X4*X8+2*X4*X6+2*X3*X7-2*X10*X5+X9*X8+2*X9*X6)+PH**2
     &  *(X1*X10-X3*X7-X4*X6)+2*(-X1*X9*X8-X2*X10*X5+X2*X9*X6+X4*
     &  X8*X5)
        FM(9,10)=0.5D0*(FMXX+FM(9,10))
        FM(10,10)=-4*PQ**4*X10-PQ**2*PH**2*X10+4*PQ**2*(-X1*X10-X2*X10+
     &  X3*X7+X4*X6-X10*X5+X9*X3+X8*X4)+PH**2*(X1*X10-X3*X7-X4*X6
     &  )+2*X5*(-X10*X2+X9*X3+X8*X4)

C...Repackage matrix elements.
        DO 200 I=1,8
          DO 190 J=I,8
            RM(I,J)=FM(I,J)
  190     CONTINUE
  200   CONTINUE
        RM(7,7)=FM(7,7)-2D0*FM(9,9)
        RM(7,8)=FM(7,8)-2D0*FM(9,10)
        RM(8,8)=FM(8,8)-2D0*FM(10,10)

C...Produce final result: matrix elements * colours * propagators.
        DO 220 I=1,8
          DO 210 J=I,8
            FAC=8D0
            IF(I.EQ.J)FAC=4D0
            WTQQBH=WTQQBH+RM(I,J)*FAC*CLR(I,J)/(DX(I)*DX(J))
  210     CONTINUE
  220   CONTINUE
        WTQQBH=-WTQQBH/256D0

      ELSE
C...Evaluate matrix elements for q + qbar -> Q + Qbar + H.
        A11=-8D0*PQ**4*X10-2D0*PQ**2*PH**2*X10-(8D0*PQ**2)*(X2*X10+X3
     &  *X7+X4*X6+X9*X6+X8*X7)+2D0*PH**2*(X3*X7+X4*X6)-(4D0*X2)*(X9
     &  *X6+X8*X7)
        A12=-8D0*PQ**4*X10+4D0*PQ**2*(-X2*X10-X3*X9-2D0*X3*X7-X4*X8-
     &  2D0*X4*X6-X10*X5-X9*X8-X9*X6-X8*X7)+2D0*PH**2*(-X1*X10+X3*X7
     &  +X4*X6)+2D0*(2D0*X1*X9*X8-X2*X9*X6-X2*X8*X7-X3*X9*X5-X4*X8*
     &  X5)
        A22=-8D0*PQ**4*X10-2D0*PQ**2*PH**2*X10-(8D0*PQ**2)*(X3*X9+X3*
     &  X7+X4*X8+X4*X6+X10*X5)+2D0*PH**2*(X3*X7+X4*X6)-(4D0*X5)*(X3
     &  *X9+X4*X8)

C...Produce final result: matrix elements * propagators.
        A11=A11/DX(7)**2
        A12=A12/(DX(7)*DX(8))
        A22=A22/DX(8)**2
        WTQQBH=-(A11+A22+2D0*A12)*8D0/9D0
      ENDIF

      RETURN
      END

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[juergen.reuter at desy dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #6 from Jürgen Reuter <juergen.reuter at desy dot de> ---
Created attachment 51716
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51716&action=edit
gfortran appearance of the ICE

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #7 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Yes, I've posted a patch for this already
https://gcc.gnu.org/pipermail/gcc-patches/2021-November/583008.html

waiting review.

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[juergen.reuter at desy dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #8 from Jürgen Reuter <juergen.reuter at desy dot de> ---
(In reply to Jürgen Reuter from comment #6)
> Created attachment 51716 [details]
> gfortran appearance of the ICE

Just for completeness, this example needs to be compiled with -O2, while
-O0 and -O1 work fine.

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[mario-baumann at web dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Mario Baumann <mario-baumann at web dot de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mario-baumann at web dot de

--- Comment #9 from Mario Baumann <mario-baumann at web dot de> ---
Created attachment 51718
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51718&action=edit
smaller Fortran example

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[mario-baumann at web dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #10 from Mario Baumann <mario-baumann at web dot de> ---
Added smaller Fortran example foo.F.
Same behaviour, fails with ICE for -O2 and works with -O1 and -O0

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tamar Christina <tnfchris@gcc.gnu.org>:

https://gcc.gnu.org/g:6cc8aa65fdeaefe9774d5e0d4e72c91f52313be1

commit r12-4846-g6cc8aa65fdeaefe9774d5e0d4e72c91f52313be1
Author: Tamar Christina <tamar.christina@arm.com>
Date:   Tue Nov 2 13:26:36 2021 +0000

    middle-end: Fix PR103007, add missing check on complex fms detection.

    The complex FMS detection is missing a check on if the nodes of the
VEC_PERM
    has the amount of children we expect before it recurses.

    This check is there on MUL and FMA but was missing for FMS, due to this the
    compiler goes on further than it should and hits an assert.

    gcc/ChangeLog:

            PR tree-optimization/103007
            * tree-vect-slp-patterns.c (complex_fms_pattern::matches): Add elem
            check.

    gcc/testsuite/ChangeLog:

            PR tree-optimization/103007
            * g++.dg/pr103007.C: New test.

[Bug tree-optimization/103007] [12 Regression] ice in vect_normalize_conj_loc, at tree-vect-slp-patterns.c:722 since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103007

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #12 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Fixed, thanks for the report.