pclarf.f

Go to the documentation of this file.

      SUBROUTINE pclarf( SIDE, M, N, V, IV, JV, DESCV, INCV, TAU,
     $                   C, IC, JC, DESCC, WORK )
*
*  -- ScaLAPACK auxiliary routine (version 1.7) --
*     University of Tennessee, Knoxville, Oak Ridge National Laboratory,
*     and University of California, Berkeley.
*     May 25, 2001
*
*     .. Scalar Arguments ..
      CHARACTER          SIDE
      INTEGER            IC, INCV, IV, JC, JV, M, N
*     ..
*     .. Array Arguments ..
      INTEGER            DESCC( * ), DESCV( * )
      COMPLEX            C( * ), TAU( * ), V( * ), WORK( * )
*     ..
*
*  Purpose
*  =======
*
*  PCLARF applies a complex elementary reflector Q to a complex M-by-N
*  distributed matrix sub( C ) = C(IC:IC+M-1,JC:JC+N-1), from either the
*  left or the right. Q is represented in the form
*
*        Q = I - tau * v * v'
*
*  where tau is a complex scalar and v is a complex vector.
*
*  If tau = 0, then Q is taken to be the unit matrix.
*
*  Notes
*  =====
*
*  Each global data object is described by an associated description
*  vector.  This vector stores the information required to establish
*  the mapping between an object element and its corresponding process
*  and memory location.
*
*  Let A be a generic term for any 2D block cyclicly distributed array.
*  Such a global array has an associated description vector DESCA.
*  In the following comments, the character _ should be read as
*  "of the global array".
*
*  NOTATION        STORED IN      EXPLANATION
*  --------------- -------------- --------------------------------------
*  DTYPE_A(global) DESCA( DTYPE_ )The descriptor type.  In this case,
*                                 DTYPE_A = 1.
*  CTXT_A (global) DESCA( CTXT_ ) The BLACS context handle, indicating
*                                 the BLACS process grid A is distribu-
*                                 ted over. The context itself is glo-
*                                 bal, but the handle (the integer
*                                 value) may vary.
*  M_A    (global) DESCA( M_ )    The number of rows in the global
*                                 array A.
*  N_A    (global) DESCA( N_ )    The number of columns in the global
*                                 array A.
*  MB_A   (global) DESCA( MB_ )   The blocking factor used to distribute
*                                 the rows of the array.
*  NB_A   (global) DESCA( NB_ )   The blocking factor used to distribute
*                                 the columns of the array.
*  RSRC_A (global) DESCA( RSRC_ ) The process row over which the first
*                                 row of the array A is distributed.
*  CSRC_A (global) DESCA( CSRC_ ) The process column over which the
*                                 first column of the array A is
*                                 distributed.
*  LLD_A  (local)  DESCA( LLD_ )  The leading dimension of the local
*                                 array.  LLD_A >= MAX(1,LOCr(M_A)).
*
*  Let K be the number of rows or columns of a distributed matrix,
*  and assume that its process grid has dimension p x q.
*  LOCr( K ) denotes the number of elements of K that a process
*  would receive if K were distributed over the p processes of its
*  process column.
*  Similarly, LOCc( K ) denotes the number of elements of K that a
*  process would receive if K were distributed over the q processes of
*  its process row.
*  The values of LOCr() and LOCc() may be determined via a call to the
*  ScaLAPACK tool function, NUMROC:
*          LOCr( M ) = NUMROC( M, MB_A, MYROW, RSRC_A, NPROW ),
*          LOCc( N ) = NUMROC( N, NB_A, MYCOL, CSRC_A, NPCOL ).
*  An upper bound for these quantities may be computed by:
*          LOCr( M ) <= ceil( ceil(M/MB_A)/NPROW )*MB_A
*          LOCc( N ) <= ceil( ceil(N/NB_A)/NPCOL )*NB_A
*
*  Because vectors may be viewed as a subclass of matrices, a
*  distributed vector is considered to be a distributed matrix.
*
*  Restrictions
*  ============
*
*  If SIDE = 'Left' and INCV = 1, then the row process having the first
*  entry V(IV,JV) must also have the first row of sub( C ). Moreover,
*  MOD(IV-1,MB_V) must be equal to MOD(IC-1,MB_C), if INCV=M_V, only
*  the last equality must be satisfied.
*
*  If SIDE = 'Right' and INCV = M_V then the column process having the
*  first entry V(IV,JV) must also have the first column of sub( C ) and
*  MOD(JV-1,NB_V) must be equal to MOD(JC-1,NB_C), if INCV = 1 only the
*  last equality must be satisfied.
*
*  Arguments
*  =========
*
*  SIDE    (global input) CHARACTER
*          = 'L': form  Q * sub( C ),
*          = 'R': form  sub( C ) * Q.
*
*  M       (global input) INTEGER
*          The number of rows to be operated on i.e the number of rows
*          of the distributed submatrix sub( C ). M >= 0.
*
*  N       (global input) INTEGER
*          The number of columns to be operated on i.e the number of
*          columns of the distributed submatrix sub( C ). N >= 0.
*
*  V       (local input) COMPLEX pointer into the local memory
*          to an array of dimension (LLD_V,*) containing the local
*          pieces of the distributed vectors V representing the
*          Householder transformation Q,
*             V(IV:IV+M-1,JV) if SIDE = 'L' and INCV = 1,
*             V(IV,JV:JV+M-1) if SIDE = 'L' and INCV = M_V,
*             V(IV:IV+N-1,JV) if SIDE = 'R' and INCV = 1,
*             V(IV,JV:JV+N-1) if SIDE = 'R' and INCV = M_V,
*
*          The vector v in the representation of Q. V is not used if
*          TAU = 0.
*
*  IV      (global input) INTEGER
*          The row index in the global array V indicating the first
*          row of sub( V ).
*
*  JV      (global input) INTEGER
*          The column index in the global array V indicating the
*          first column of sub( V ).
*
*  DESCV   (global and local input) INTEGER array of dimension DLEN_.
*          The array descriptor for the distributed matrix V.
*
*  INCV    (global input) INTEGER
*          The global increment for the elements of V. Only two values
*          of INCV are supported in this version, namely 1 and M_V.
*          INCV must not be zero.
*
*  TAU     (local input) COMPLEX, array, dimension  LOCc(JV) if
*          INCV = 1, and LOCr(IV) otherwise. This array contains the
*          Householder scalars related to the Householder vectors.
*          TAU is tied to the distributed matrix V.
*
*  C       (local input/local output) COMPLEX pointer into the
*          local memory to an array of dimension (LLD_C, LOCc(JC+N-1) ),
*          containing the local pieces of sub( C ). On exit, sub( C )
*          is overwritten by the Q * sub( C ) if SIDE = 'L', or
*          sub( C ) * Q if SIDE = 'R'.
*
*  IC      (global input) INTEGER
*          The row index in the global array C indicating the first
*          row of sub( C ).
*
*  JC      (global input) INTEGER
*          The column index in the global array C indicating the
*          first column of sub( C ).
*
*  DESCC   (global and local input) INTEGER array of dimension DLEN_.
*          The array descriptor for the distributed matrix C.
*
*  WORK    (local workspace) COMPLEX array, dimension (LWORK)
*          If INCV = 1,
*            if SIDE = 'L',
*              if IVCOL = ICCOL,
*                LWORK >= NqC0
*              else
*                LWORK >= MpC0 + MAX( 1, NqC0 )
*              end if
*            else if SIDE = 'R',
*              LWORK >= NqC0 + MAX( MAX( 1, MpC0 ), NUMROC( NUMROC(
*                       N+ICOFFC,NB_V,0,0,NPCOL ),NB_V,0,0,LCMQ ) )
*            end if
*          else if INCV = M_V,
*            if SIDE = 'L',
*              LWORK >= MpC0 + MAX( MAX( 1, NqC0 ), NUMROC( NUMROC(
*                       M+IROFFC,MB_V,0,0,NPROW ),MB_V,0,0,LCMP ) )
*            else if SIDE = 'R',
*              if IVROW = ICROW,
*                LWORK >= MpC0
*              else
*                LWORK >= NqC0 + MAX( 1, MpC0 )
*              end if
*            end if
*          end if
*
*          where LCM is the least common multiple of NPROW and NPCOL and
*          LCM = ILCM( NPROW, NPCOL ), LCMP = LCM / NPROW,
*          LCMQ = LCM / NPCOL,
*
*          IROFFC = MOD( IC-1, MB_C ), ICOFFC = MOD( JC-1, NB_C ),
*          ICROW = INDXG2P( IC, MB_C, MYROW, RSRC_C, NPROW ),
*          ICCOL = INDXG2P( JC, NB_C, MYCOL, CSRC_C, NPCOL ),
*          MpC0 = NUMROC( M+IROFFC, MB_C, MYROW, ICROW, NPROW ),
*          NqC0 = NUMROC( N+ICOFFC, NB_C, MYCOL, ICCOL, NPCOL ),
*
*          ILCM, INDXG2P and NUMROC are ScaLAPACK tool functions;
*          MYROW, MYCOL, NPROW and NPCOL can be determined by calling
*          the subroutine BLACS_GRIDINFO.
*
*  Alignment requirements
*  ======================
*
*  The distributed submatrices V(IV:*, JV:*) and C(IC:IC+M-1,JC:JC+N-1)
*  must verify some alignment properties, namely the following
*  expressions should be true:
*
*  MB_V = NB_V,
*
*  If INCV = 1,
*    If SIDE = 'Left',
*      ( MB_V.EQ.MB_C .AND. IROFFV.EQ.IROFFC .AND. IVROW.EQ.ICROW )
*    If SIDE = 'Right',
*      ( MB_V.EQ.NB_A .AND. MB_V.EQ.NB_C .AND. IROFFV.EQ.ICOFFC )
*  else if INCV = M_V,
*    If SIDE = 'Left',
*      ( MB_V.EQ.NB_V .AND. MB_V.EQ.MB_C .AND. ICOFFV.EQ.IROFFC )
*    If SIDE = 'Right',
*      ( NB_V.EQ.NB_C .AND. ICOFFV.EQ.ICOFFC .AND. IVCOL.EQ.ICCOL )
*  end if
*
*  =====================================================================
*
*     .. Parameters ..
      INTEGER            BLOCK_CYCLIC_2D, CSRC_, CTXT_, DLEN_, DTYPE_,
     $                   lld_, mb_, m_, nb_, n_, rsrc_
      parameter( block_cyclic_2d = 1, dlen_ = 9, dtype_ = 1,
     $                     ctxt_ = 2, m_ = 3, n_ = 4, mb_ = 5, nb_ = 6,
     $                     rsrc_ = 7, csrc_ = 8, lld_ = 9 )
      COMPLEX            ONE, ZERO
      parameter( one  = ( 1.0e+0, 0.0e+0 ),
     $                     zero = ( 0.0e+0, 0.0e+0 ) )
*     ..
*     .. Local Scalars ..
      LOGICAL            CCBLCK, CRBLCK
      CHARACTER          COLBTOP, ROWBTOP
      INTEGER            ICCOL, ICOFF, ICROW, ICTXT, IIC, IIV, IOFFC,
     $                   ioffv, ipw, iroff, ivcol, ivrow, jjc, jjv, ldc,
     $                   ldv, mycol, myrow, mp, ncc, ncv, npcol, nprow,
     $                   nq, rdest
      COMPLEX            TAULOC( 1 )
*     ..
*     .. External Subroutines ..
      EXTERNAL           blacs_gridinfo, ccopy, cgebr2d, cgebs2d,
     $                   cgemv, cgerc, cgerv2d, cgesd2d,
     $                   cgsum2d, claset, infog2l, pb_topget,
     $                   pbctrnv
*     ..
*     .. External Functions ..
      LOGICAL            LSAME
      INTEGER            NUMROC
      EXTERNAL           lsame, numroc
*     ..
*     .. Intrinsic Functions ..
      INTRINSIC          min, mod
*     ..
*     .. Executable Statements ..
*
*     Quick return if possible
*
      IF( m.LE.0 .OR. n.LE.0 )
     $   RETURN
*
*     Get grid parameters.
*
      ictxt = descc( ctxt_ )
      CALL blacs_gridinfo( ictxt, nprow, npcol, myrow, mycol )
*
*     Figure local indexes
*
      CALL infog2l( ic, jc, descc, nprow, npcol, myrow, mycol, iic, jjc,
     $              icrow, iccol )
      CALL infog2l( iv, jv, descv, nprow, npcol, myrow, mycol, iiv, jjv,
     $              ivrow, ivcol )
      ncc = numroc( descc( n_ ), descc( nb_ ), mycol, descc( csrc_ ),
     $              npcol )
      ncv = numroc( descv( n_ ), descv( nb_ ), mycol, descv( csrc_ ),
     $              npcol )
      ldc = descc( lld_ )
      ldv = descv( lld_ )
      iic = min( iic, ldc )
      iiv = min( iiv, ldv )
      jjc = min( jjc, ncc )
      jjv = min( jjv, ncv )
      ioffc = iic+(jjc-1)*ldc
      ioffv = iiv+(jjv-1)*ldv
*
      iroff = mod( ic-1, descc( mb_ ) )
      icoff = mod( jc-1, descc( nb_ ) )
      mp = numroc( m+iroff, descc( mb_ ), myrow, icrow, nprow )
      nq = numroc( n+icoff, descc( nb_ ), mycol, iccol, npcol )
      IF( myrow.EQ.icrow )
     $   mp = mp - iroff
      IF( mycol.EQ.iccol )
     $   nq = nq - icoff
*
*     Is sub( C ) only distributed over a process row ?
*
      crblck = ( m.LE.(descc( mb_ )-iroff) )
*
*     Is sub( C ) only distributed over a process column ?
*
      ccblck = ( n.LE.(descc( nb_ )-icoff) )
*
      IF( lsame( side, 'L' ) ) THEN
*
         IF( crblck ) THEN
            rdest = icrow
         ELSE
            rdest = -1
         END IF
*
         IF( ccblck ) THEN
*
*           sub( C ) is distributed over a process column
*
            IF( descv( m_ ).EQ.incv ) THEN
*
*              Transpose row vector V
*
               ipw = mp+1
               CALL pbctrnv( ictxt, 'Rowwise', 'Transpose', m,
     $                       descv( nb_ ), iroff, v( ioffv ), ldv, zero,
     $                       work, 1, ivrow, ivcol, icrow, iccol,
     $                       work( ipw ) )
*
*              Perform the local computation within a process column
*
               IF( mycol.EQ.iccol ) THEN
*
                  IF( myrow.EQ.ivrow ) THEN
*
                     CALL cgebs2d( ictxt, 'Columnwise', ' ', 1, 1,
     $                             tau( iiv ), 1 )
                     tauloc( 1 ) = tau( iiv )
*
                  ELSE
*
                     CALL cgebr2d( ictxt, 'Columnwise', ' ', 1, 1,
     $                             tauloc, 1, ivrow, mycol )
*
                  END IF
*
                  IF( tauloc( 1 ).NE.zero ) THEN
*
*                    w := sub( C )' * v
*
                     IF( mp.GT.0 ) THEN
                        CALL cgemv( 'Conjugate transpose', mp, nq, one,
     $                              c( ioffc ), ldc, work, 1, zero,
     $                              work( ipw ), 1 )
                     ELSE
                        CALL claset( 'All', nq, 1, zero, zero,
     $                               work( ipw ), max( 1, nq ) )
                     END IF
                     CALL cgsum2d( ictxt, 'Columnwise', ' ', nq, 1,
     $                             work( ipw ), max( 1, nq ), rdest,
     $                             mycol )
*
*                    sub( C ) := sub( C ) - v * w'
*
                     CALL cgerc( mp, nq, -tauloc( 1 ), work, 1,
     $                           work( ipw ), 1, c( ioffc ), ldc )
                  END IF
*
               END IF
*
            ELSE
*
*              V is a column vector
*
               IF( ivcol.EQ.iccol ) THEN
*
*                 Perform the local computation within a process column
*
                  IF( mycol.EQ.iccol ) THEN
*
                     tauloc( 1 ) = tau( jjv )
*
                     IF( tauloc( 1 ).NE.zero ) THEN
*
*                       w := sub( C )' * v
*
                        IF( mp.GT.0 ) THEN
                           CALL cgemv( 'Conjugate transpose', mp, nq,
     $                              one, c( ioffc ), ldc, v( ioffv ), 1,
     $                              zero, work, 1 )
                        ELSE
                           CALL claset( 'All', nq, 1, zero, zero,
     $                                  work, max( 1, nq ) )
                        END IF
                        CALL cgsum2d( ictxt, 'Columnwise', ' ', nq, 1,
     $                                work, max( 1, nq ), rdest, mycol )
*
*                       sub( C ) := sub( C ) - v * w'
*
                        CALL cgerc( mp, nq, -tauloc( 1 ), v( ioffv ), 1,
     $                              work, 1, c( ioffc ), ldc )
                     END IF
*
                  END IF
*
               ELSE
*
*                 Send V and TAU to the process column ICCOL
*
                  IF( mycol.EQ.ivcol ) THEN
*
                     ipw = mp+1
                     CALL ccopy( mp, v( ioffv ), 1, work, 1 )
                     work( ipw ) = tau( jjv )
                     CALL cgesd2d( ictxt, ipw, 1, work, ipw, myrow,
     $                             iccol )
*
                  ELSE IF( mycol.EQ.iccol ) THEN
*
                     ipw = mp+1
                     CALL cgerv2d( ictxt, ipw, 1, work, ipw, myrow,
     $                             ivcol )
                     tauloc( 1 ) = work( ipw )
*
                     IF( tauloc( 1 ).NE.zero ) THEN
*
*                       w := sub( C )' * v
*
                        IF( mp.GT.0 ) THEN
                           CALL cgemv( 'Conjugate transpose', mp, nq,
     $                                 one, c( ioffc ), ldc, work, 1,
     $                                 zero, work( ipw ), 1 )
                        ELSE
                           CALL claset( 'All', nq, 1, zero, zero,
     $                                  work( ipw ), max( 1, nq ) )
                        END IF
                        CALL cgsum2d( ictxt, 'Columnwise', ' ', nq, 1,
     $                                work( ipw ), max( 1, nq ), rdest,
     $                                mycol )
*
*                       sub( C ) := sub( C ) - v * w'
*
                        CALL cgerc( mp, nq, -tauloc( 1 ), work, 1,
     $                              work( ipw ), 1, c( ioffc ), ldc )
                     END IF
*
                  END IF
*
               END IF
*
            END IF
*
         ELSE
*
*           sub( C ) is a proper distributed matrix
*
            IF( descv( m_ ).EQ.incv ) THEN
*
*              Transpose and broadcast row vector V
*
               ipw = mp+1
               CALL pbctrnv( ictxt, 'Rowwise', 'transpose', M,
     $                       DESCV( NB_ ), IROFF, V( IOFFV ), LDV, ZERO,
     $                       WORK, 1, IVROW, IVCOL, ICROW, -1,
     $                       WORK( IPW ) )
*
*              Perform the local computation within a process column
*
.EQ.               IF( MYROWIVROW ) THEN
*
                  CALL CGEBS2D( ICTXT, 'columnwise', ' ', 1, 1,
     $                          TAU( IIV ), 1 )
                  TAULOC( 1 ) = TAU( IIV )
*
               ELSE
*
                  CALL CGEBR2D( ICTXT, 'columnwise', ' ', 1, 1, TAULOC,
     $                          1, IVROW, MYCOL )
*
               END IF
*
.NE.               IF( TAULOC( 1 )ZERO ) THEN
*
*                 w := sub( C )' * v
*
.GT.                  IF( MP0 ) THEN
.GT.                     IF( IOFFC0 )
     $                  CALL CGEMV( 'conjugate transpose', MP, NQ, ONE,
     $                              C( IOFFC ), LDC, WORK, 1, ZERO,
     $                              WORK( IPW ), 1 )
                  ELSE
                     CALL CLASET( 'all', NQ, 1, ZERO, ZERO,
     $                            WORK( IPW ), MAX( 1, NQ ) )
                  END IF
                  CALL CGSUM2D( ICTXT, 'columnwise', ' ', NQ, 1,
     $                          WORK( IPW ), MAX( 1, NQ ), RDEST,
     $                          MYCOL )
*
*                 sub( C ) := sub( C ) - v * w'
*
.GT.                  IF( IOFFC0 )
     $               CALL CGERC( MP, NQ, -TAULOC( 1 ), WORK, 1,
     $                           WORK( IPW ), 1, C( IOFFC ), LDC )
               END IF
*
            ELSE
*
*              Broadcast column vector V
*
               CALL PB_TOPGET( ICTXT, 'broadcast', 'rowwise', ROWBTOP )
.EQ.               IF( MYCOLIVCOL ) THEN
*
                  IPW = MP+1
                  CALL CCOPY( MP, V( IOFFV ), 1, WORK, 1 )
                  WORK(IPW) = TAU( JJV )
                  CALL CGEBS2D( ICTXT, 'rowwise', ROWBTOP, IPW, 1,
     $                          WORK, IPW )
                  TAULOC( 1 ) = TAU( JJV )
*
               ELSE
*
                  IPW = MP+1
                  CALL CGEBR2D( ICTXT, 'rowwise', ROWBTOP, IPW, 1, WORK,
     $                          IPW, MYROW, IVCOL )
                  TAULOC( 1 ) = WORK( IPW )
*
               END IF
*
.NE.               IF( TAULOC( 1 )ZERO ) THEN
*
*                 w := sub( C )' * v
*
.GT.                  IF( MP0 ) THEN
.GT.                     IF( IOFFC0 )
     $                  CALL CGEMV( 'conjugate transpose', MP, NQ, ONE,
     $                              C( IOFFC ), LDC, WORK, 1, ZERO,
     $                              WORK( IPW ), 1 )
                  ELSE
                     CALL CLASET( 'all', NQ, 1, ZERO, ZERO,
     $                            WORK( IPW ), MAX( 1, NQ ) )
                  END IF
                  CALL CGSUM2D( ICTXT, 'columnwise', ' ', NQ, 1,
     $                          WORK( IPW ), MAX( 1, NQ ), RDEST,
     $                          MYCOL )
*
*                 sub( C ) := sub( C ) - v * w'
*
.GT.                  IF( IOFFC0 )
     $               CALL CGERC( MP, NQ, -TAULOC( 1 ), WORK, 1,
     $                           WORK( IPW ), 1, C( IOFFC ), LDC )
               END IF
*
            END IF
*
         END IF
*
      ELSE
*
         IF( CCBLCK ) THEN
            RDEST = MYROW
         ELSE
            RDEST = -1
         END IF
*
         IF( CRBLCK ) THEN
*
*           sub( C ) is distributed over a process row
*
.EQ.            IF( DESCV( M_ )INCV ) THEN
*
*              V is a row vector
*
.EQ.               IF( IVROWICROW ) THEN
*
*                 Perform the local computation within a process row
*
.EQ.                  IF( MYROWICROW ) THEN
*
                     TAULOC( 1 ) = TAU( IIV )
*
.NE.                     IF( TAULOC( 1 )ZERO ) THEN
*
*                       w := sub( C ) * v
*
.GT.                        IF( NQ0 ) THEN
                           CALL CGEMV( 'no transpose', MP, NQ, ONE,
     $                                 C( IOFFC ), LDC, V( IOFFV ), LDV,
     $                                 ZERO, WORK, 1 )
                        ELSE
                           CALL CLASET( 'all', mp, 1, zero, zero,
     $                                  work, max( 1, mp ) )
                        END IF
                        CALL cgsum2d( ictxt, 'Rowwise', ' ', mp, 1,
     $                                work, max( 1, mp ), rdest, iccol )
*
*                       sub( C ) := sub( C ) - w * v'
*
                        IF( ioffv.GT.0 .AND. ioffc.GT.0 )
     $                     CALL cgerc( mp, nq, -tauloc( 1 ), work, 1,
     $                                 v( ioffv ), ldv, c( ioffc ),
     $                                 ldc )
                     END IF
*
                  END IF
*
               ELSE
*
*                 Send V and TAU to the process row ICROW
*
                  IF( myrow.EQ.ivrow ) THEN
*
                     ipw = nq+1
                     CALL ccopy( nq, v( ioffv ), ldv, work, 1 )
                     work(ipw) = tau( iiv )
                     CALL cgesd2d( ictxt, ipw, 1, work, ipw, icrow,
     $                             mycol )
*
                  ELSE IF( myrow.EQ.icrow ) THEN
*
                     ipw = nq+1
                     CALL cgerv2d( ictxt, ipw, 1, work, ipw, ivrow,
     $                             mycol )
                     tauloc( 1 ) = work( ipw )
*
                     IF( tauloc( 1 ).NE.zero ) THEN
*
*                       w := sub( C ) * v
*
                        IF( nq.GT.0 ) THEN
                           CALL cgemv( 'No transpose', mp, nq, one,
     $                                 c( ioffc ), ldc, work, 1, zero,
     $                                 work( ipw ), 1 )
                        ELSE
                           CALL claset( 'All', mp, 1, zero, zero,
     $                                  work( ipw ), max( 1, mp ) )
                        END IF
                        CALL cgsum2d( ictxt, 'Rowwise', ' ', mp, 1,
     $                                work( ipw ), max( 1, mp ), rdest,
     $                                iccol )
*
*                       sub( C ) := sub( C ) - w * v'
*
                        CALL cgerc( mp, nq, -tauloc( 1 ), work( ipw ),
     $                              1, work, 1, c( ioffc ), ldc )
                     END IF
*
                  END IF
*
               END IF
*
            ELSE
*
*              Transpose column vector V
*
               ipw = nq+1
               CALL pbctrnv( ictxt, 'Columnwise', 'Transpose', n,
     $                       descv( mb_ ), icoff, v( ioffv ), 1, zero,
     $                       work, 1, ivrow, ivcol, icrow, iccol,
     $                       work( ipw ) )
*
*              Perform the local computation within a process column
*
               IF( myrow.EQ.icrow ) THEN
*
                  IF( mycol.EQ.ivcol ) THEN
*
                     CALL cgebs2d( ictxt, 'Rowwise', ' ', 1, 1,
     $                             tau( jjv ), 1 )
                     tauloc( 1 ) = tau( jjv )
*
                  ELSE
*
                     CALL cgebr2d( ictxt, 'Rowwise', ' ', 1, 1, tauloc,
     $                             1, myrow, ivcol )
*
                  END IF
*
                  IF( tauloc( 1 ).NE.zero ) THEN
*
*                    w := sub( C ) * v
*
                     IF( nq.GT.0 ) THEN
                        CALL cgemv( 'No transpose', mp, nq, one,
     $                              c( ioffc ), ldc, work, 1, zero,
     $                              work( ipw ), 1 )
                     ELSE
                        CALL claset( 'All', mp, 1, zero, zero,
     $                               work( ipw ), max( 1, mp ) )
                     END IF
                     CALL cgsum2d( ictxt, 'Rowwise', ' ', mp, 1,
     $                             work( ipw ), max( 1, mp ), rdest,
     $                             iccol )
*
*                    sub( C ) := sub( C ) - w * v'
*
                     CALL cgerc( mp, nq, -tauloc( 1 ), work( ipw ), 1,
     $                           work, 1, c( ioffc ), ldc )
                  END IF
*
               END IF
*
            END IF
*
         ELSE
*
*           sub( C ) is a proper distributed matrix
*
            IF( descv( m_ ).EQ.incv ) THEN
*
*              Broadcast row vector V
*
               CALL pb_topget( ictxt, 'Broadcast', 'Columnwise',
     $                         colbtop )
               IF( myrow.EQ.ivrow ) THEN
*
                  ipw = nq+1
                  IF( ioffv.GT.0 )
     $               CALL ccopy( nq, v( ioffv ), ldv, work, 1 )
                  work(ipw) = tau( iiv )
                  CALL cgebs2d( ictxt, 'Columnwise', colbtop, ipw, 1,
     $                          work, ipw )
                  tauloc( 1 ) = tau( iiv )
*
               ELSE
*
                  ipw = nq+1
                  CALL cgebr2d( ictxt, 'Columnwise', colbtop, ipw, 1,
     $                          work, ipw, ivrow, mycol )
                  tauloc( 1 ) = work( ipw )
*
               END IF
*
               IF( tauloc( 1 ).NE.zero ) THEN
*
*                 w := sub( C ) * v
*
                  IF( nq.GT.0 ) THEN
                     CALL cgemv( 'No Transpose', mp, nq, one,
     $                           c( ioffc ), ldc, work, 1, zero,
     $                           work( ipw ), 1 )
                  ELSE
                     CALL claset( 'All', mp, 1, zero, zero,
     $                            work( ipw ), max( 1, mp ) )
                  END IF
                  CALL cgsum2d( ictxt, 'Rowwise', ' ', mp, 1,
     $                          work( ipw ), max( 1, mp ), rdest,
     $                          iccol )
*
*                 sub( C ) := sub( C ) - w * v'
*
                  IF( ioffc.GT.0 )
     $               CALL cgerc( mp, nq, -tauloc( 1 ), work( ipw ), 1,
     $                           work, 1, c( ioffc ), ldc )
               END IF
*
            ELSE
*
*              Transpose and broadcast column vector V
*
               ipw = nq+1
               CALL pbctrnv( ictxt, 'Columnwise', 'Transpose', n,
     $                       descv( mb_ ), icoff, v( ioffv ), 1, zero,
     $                       work, 1, ivrow, ivcol, -1, iccol,
     $                       work( ipw ) )
*
*              Perform the local computation within a process column
*
               IF( mycol.EQ.ivcol ) THEN
*
                  CALL cgebs2d( ictxt, 'Rowwise', ' ', 1, 1, tau( jjv ),
     $                          1 )
                  tauloc( 1 ) = tau( jjv )
*
               ELSE
*
                  CALL cgebr2d( ictxt, 'Rowwise', ' ', 1, 1, tauloc, 1,
     $                          myrow, ivcol )
*
               END IF
*
               IF( tauloc( 1 ).NE.zero ) THEN
*
*                 w := sub( C ) * v
*
                  IF( nq.GT.0 ) THEN
                     CALL cgemv( 'No transpose', mp, nq, one,
     $                           c( ioffc ), ldc, work, 1, zero,
     $                           work( ipw ), 1 )
                  ELSE
                     CALL claset( 'All', mp, 1, zero, zero, work( ipw ),
     $                            max( 1, mp ) )
                  END IF
                  CALL cgsum2d( ictxt, 'Rowwise', ' ', mp, 1,
     $                          work( ipw ), max( 1, mp ), rdest,
     $                          iccol )
*
*                 sub( C ) := sub( C ) - w * v'
*
                  CALL cgerc( mp, nq, -tauloc( 1 ), work( ipw ), 1,
     $                        work, 1, c( ioffc ), ldc )
               END IF
*
            END IF
*
         END IF
*
      END IF
*
      RETURN
*
*     End of PCLARF
*
      END