File:  [local] / rpl / lapack / lapack / zunmrz.f
Revision 1.1.1.1 (vendor branch): download - view: text, annotated - select for diffs - revision graph
Tue Jan 26 15:22:45 2010 UTC (14 years, 3 months ago) by bertrand
Branches: JKB
CVS tags: start, rpl-4_0_14, rpl-4_0_13, rpl-4_0_12, rpl-4_0_11, rpl-4_0_10


Commit initial.

    1:       SUBROUTINE ZUNMRZ( SIDE, TRANS, M, N, K, L, A, LDA, TAU, C, LDC,
    2:      $                   WORK, LWORK, INFO )
    3: *
    4: *  -- LAPACK routine (version 3.2) --
    5: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
    6: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
    7: *     January 2007
    8: *
    9: *     .. Scalar Arguments ..
   10:       CHARACTER          SIDE, TRANS
   11:       INTEGER            INFO, K, L, LDA, LDC, LWORK, M, N
   12: *     ..
   13: *     .. Array Arguments ..
   14:       COMPLEX*16         A( LDA, * ), C( LDC, * ), TAU( * ), WORK( * )
   15: *     ..
   16: *
   17: *  Purpose
   18: *  =======
   19: *
   20: *  ZUNMRZ overwrites the general complex M-by-N matrix C with
   21: *
   22: *                  SIDE = 'L'     SIDE = 'R'
   23: *  TRANS = 'N':      Q * C          C * Q
   24: *  TRANS = 'C':      Q**H * C       C * Q**H
   25: *
   26: *  where Q is a complex unitary matrix defined as the product of k
   27: *  elementary reflectors
   28: *
   29: *        Q = H(1) H(2) . . . H(k)
   30: *
   31: *  as returned by ZTZRZF. Q is of order M if SIDE = 'L' and of order N
   32: *  if SIDE = 'R'.
   33: *
   34: *  Arguments
   35: *  =========
   36: *
   37: *  SIDE    (input) CHARACTER*1
   38: *          = 'L': apply Q or Q**H from the Left;
   39: *          = 'R': apply Q or Q**H from the Right.
   40: *
   41: *  TRANS   (input) CHARACTER*1
   42: *          = 'N':  No transpose, apply Q;
   43: *          = 'C':  Conjugate transpose, apply Q**H.
   44: *
   45: *  M       (input) INTEGER
   46: *          The number of rows of the matrix C. M >= 0.
   47: *
   48: *  N       (input) INTEGER
   49: *          The number of columns of the matrix C. N >= 0.
   50: *
   51: *  K       (input) INTEGER
   52: *          The number of elementary reflectors whose product defines
   53: *          the matrix Q.
   54: *          If SIDE = 'L', M >= K >= 0;
   55: *          if SIDE = 'R', N >= K >= 0.
   56: *
   57: *  L       (input) INTEGER
   58: *          The number of columns of the matrix A containing
   59: *          the meaningful part of the Householder reflectors.
   60: *          If SIDE = 'L', M >= L >= 0, if SIDE = 'R', N >= L >= 0.
   61: *
   62: *  A       (input) COMPLEX*16 array, dimension
   63: *                               (LDA,M) if SIDE = 'L',
   64: *                               (LDA,N) if SIDE = 'R'
   65: *          The i-th row must contain the vector which defines the
   66: *          elementary reflector H(i), for i = 1,2,...,k, as returned by
   67: *          ZTZRZF in the last k rows of its array argument A.
   68: *          A is modified by the routine but restored on exit.
   69: *
   70: *  LDA     (input) INTEGER
   71: *          The leading dimension of the array A. LDA >= max(1,K).
   72: *
   73: *  TAU     (input) COMPLEX*16 array, dimension (K)
   74: *          TAU(i) must contain the scalar factor of the elementary
   75: *          reflector H(i), as returned by ZTZRZF.
   76: *
   77: *  C       (input/output) COMPLEX*16 array, dimension (LDC,N)
   78: *          On entry, the M-by-N matrix C.
   79: *          On exit, C is overwritten by Q*C or Q**H*C or C*Q**H or C*Q.
   80: *
   81: *  LDC     (input) INTEGER
   82: *          The leading dimension of the array C. LDC >= max(1,M).
   83: *
   84: *  WORK    (workspace/output) COMPLEX*16 array, dimension (MAX(1,LWORK))
   85: *          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
   86: *
   87: *  LWORK   (input) INTEGER
   88: *          The dimension of the array WORK.
   89: *          If SIDE = 'L', LWORK >= max(1,N);
   90: *          if SIDE = 'R', LWORK >= max(1,M).
   91: *          For optimum performance LWORK >= N*NB if SIDE = 'L', and
   92: *          LWORK >= M*NB if SIDE = 'R', where NB is the optimal
   93: *          blocksize.
   94: *
   95: *          If LWORK = -1, then a workspace query is assumed; the routine
   96: *          only calculates the optimal size of the WORK array, returns
   97: *          this value as the first entry of the WORK array, and no error
   98: *          message related to LWORK is issued by XERBLA.
   99: *
  100: *  INFO    (output) INTEGER
  101: *          = 0:  successful exit
  102: *          < 0:  if INFO = -i, the i-th argument had an illegal value
  103: *
  104: *  Further Details
  105: *  ===============
  106: *
  107: *  Based on contributions by
  108: *    A. Petitet, Computer Science Dept., Univ. of Tenn., Knoxville, USA
  109: *
  110: *  =====================================================================
  111: *
  112: *     .. Parameters ..
  113:       INTEGER            NBMAX, LDT
  114:       PARAMETER          ( NBMAX = 64, LDT = NBMAX+1 )
  115: *     ..
  116: *     .. Local Scalars ..
  117:       LOGICAL            LEFT, LQUERY, NOTRAN
  118:       CHARACTER          TRANST
  119:       INTEGER            I, I1, I2, I3, IB, IC, IINFO, IWS, JA, JC,
  120:      $                   LDWORK, LWKOPT, MI, NB, NBMIN, NI, NQ, NW
  121: *     ..
  122: *     .. Local Arrays ..
  123:       COMPLEX*16         T( LDT, NBMAX )
  124: *     ..
  125: *     .. External Functions ..
  126:       LOGICAL            LSAME
  127:       INTEGER            ILAENV
  128:       EXTERNAL           LSAME, ILAENV
  129: *     ..
  130: *     .. External Subroutines ..
  131:       EXTERNAL           XERBLA, ZLARZB, ZLARZT, ZUNMR3
  132: *     ..
  133: *     .. Intrinsic Functions ..
  134:       INTRINSIC          MAX, MIN
  135: *     ..
  136: *     .. Executable Statements ..
  137: *
  138: *     Test the input arguments
  139: *
  140:       INFO = 0
  141:       LEFT = LSAME( SIDE, 'L' )
  142:       NOTRAN = LSAME( TRANS, 'N' )
  143:       LQUERY = ( LWORK.EQ.-1 )
  144: *
  145: *     NQ is the order of Q and NW is the minimum dimension of WORK
  146: *
  147:       IF( LEFT ) THEN
  148:          NQ = M
  149:          NW = MAX( 1, N )
  150:       ELSE
  151:          NQ = N
  152:          NW = MAX( 1, M )
  153:       END IF
  154:       IF( .NOT.LEFT .AND. .NOT.LSAME( SIDE, 'R' ) ) THEN
  155:          INFO = -1
  156:       ELSE IF( .NOT.NOTRAN .AND. .NOT.LSAME( TRANS, 'C' ) ) THEN
  157:          INFO = -2
  158:       ELSE IF( M.LT.0 ) THEN
  159:          INFO = -3
  160:       ELSE IF( N.LT.0 ) THEN
  161:          INFO = -4
  162:       ELSE IF( K.LT.0 .OR. K.GT.NQ ) THEN
  163:          INFO = -5
  164:       ELSE IF( L.LT.0 .OR. ( LEFT .AND. ( L.GT.M ) ) .OR.
  165:      $         ( .NOT.LEFT .AND. ( L.GT.N ) ) ) THEN
  166:          INFO = -6
  167:       ELSE IF( LDA.LT.MAX( 1, K ) ) THEN
  168:          INFO = -8
  169:       ELSE IF( LDC.LT.MAX( 1, M ) ) THEN
  170:          INFO = -11
  171:       END IF
  172: *
  173:       IF( INFO.EQ.0 ) THEN
  174:          IF( M.EQ.0 .OR. N.EQ.0 ) THEN
  175:             LWKOPT = 1
  176:          ELSE
  177: *
  178: *           Determine the block size.  NB may be at most NBMAX, where
  179: *           NBMAX is used to define the local array T.
  180: *
  181:             NB = MIN( NBMAX, ILAENV( 1, 'ZUNMRQ', SIDE // TRANS, M, N,
  182:      $                               K, -1 ) )
  183:             LWKOPT = NW*NB
  184:          END IF
  185:          WORK( 1 ) = LWKOPT
  186: *
  187:          IF( LWORK.LT.MAX( 1, NW ) .AND. .NOT.LQUERY ) THEN
  188:             INFO = -13
  189:          END IF
  190:       END IF
  191: *
  192:       IF( INFO.NE.0 ) THEN
  193:          CALL XERBLA( 'ZUNMRZ', -INFO )
  194:          RETURN
  195:       ELSE IF( LQUERY ) THEN
  196:          RETURN
  197:       END IF
  198: *
  199: *     Quick return if possible
  200: *
  201:       IF( M.EQ.0 .OR. N.EQ.0 ) THEN
  202:          RETURN
  203:       END IF
  204: *
  205: *     Determine the block size.  NB may be at most NBMAX, where NBMAX
  206: *     is used to define the local array T.
  207: *
  208:       NB = MIN( NBMAX, ILAENV( 1, 'ZUNMRQ', SIDE // TRANS, M, N, K,
  209:      $     -1 ) )
  210:       NBMIN = 2
  211:       LDWORK = NW
  212:       IF( NB.GT.1 .AND. NB.LT.K ) THEN
  213:          IWS = NW*NB
  214:          IF( LWORK.LT.IWS ) THEN
  215:             NB = LWORK / LDWORK
  216:             NBMIN = MAX( 2, ILAENV( 2, 'ZUNMRQ', SIDE // TRANS, M, N, K,
  217:      $              -1 ) )
  218:          END IF
  219:       ELSE
  220:          IWS = NW
  221:       END IF
  222: *
  223:       IF( NB.LT.NBMIN .OR. NB.GE.K ) THEN
  224: *
  225: *        Use unblocked code
  226: *
  227:          CALL ZUNMR3( SIDE, TRANS, M, N, K, L, A, LDA, TAU, C, LDC,
  228:      $                WORK, IINFO )
  229:       ELSE
  230: *
  231: *        Use blocked code
  232: *
  233:          IF( ( LEFT .AND. .NOT.NOTRAN ) .OR.
  234:      $       ( .NOT.LEFT .AND. NOTRAN ) ) THEN
  235:             I1 = 1
  236:             I2 = K
  237:             I3 = NB
  238:          ELSE
  239:             I1 = ( ( K-1 ) / NB )*NB + 1
  240:             I2 = 1
  241:             I3 = -NB
  242:          END IF
  243: *
  244:          IF( LEFT ) THEN
  245:             NI = N
  246:             JC = 1
  247:             JA = M - L + 1
  248:          ELSE
  249:             MI = M
  250:             IC = 1
  251:             JA = N - L + 1
  252:          END IF
  253: *
  254:          IF( NOTRAN ) THEN
  255:             TRANST = 'C'
  256:          ELSE
  257:             TRANST = 'N'
  258:          END IF
  259: *
  260:          DO 10 I = I1, I2, I3
  261:             IB = MIN( NB, K-I+1 )
  262: *
  263: *           Form the triangular factor of the block reflector
  264: *           H = H(i+ib-1) . . . H(i+1) H(i)
  265: *
  266:             CALL ZLARZT( 'Backward', 'Rowwise', L, IB, A( I, JA ), LDA,
  267:      $                   TAU( I ), T, LDT )
  268: *
  269:             IF( LEFT ) THEN
  270: *
  271: *              H or H' is applied to C(i:m,1:n)
  272: *
  273:                MI = M - I + 1
  274:                IC = I
  275:             ELSE
  276: *
  277: *              H or H' is applied to C(1:m,i:n)
  278: *
  279:                NI = N - I + 1
  280:                JC = I
  281:             END IF
  282: *
  283: *           Apply H or H'
  284: *
  285:             CALL ZLARZB( SIDE, TRANST, 'Backward', 'Rowwise', MI, NI,
  286:      $                   IB, L, A( I, JA ), LDA, T, LDT, C( IC, JC ),
  287:      $                   LDC, WORK, LDWORK )
  288:    10    CONTINUE
  289: *
  290:       END IF
  291: *
  292:       WORK( 1 ) = LWKOPT
  293: *
  294:       RETURN
  295: *
  296: *     End of ZUNMRZ
  297: *
  298:       END

CVSweb interface <joel.bertrand@systella.fr>