SLATEC Routines --- DBSPVD ---

*DECK DBSPVD
      SUBROUTINE DBSPVD (T, K, NDERIV, X, ILEFT, LDVNIK, VNIKX, WORK)
C***BEGIN PROLOGUE  DBSPVD
C***PURPOSE  Calculate the value and all derivatives of order less than
C            NDERIV of all basis functions which do not vanish at X.
C***LIBRARY   SLATEC
C***CATEGORY  E3, K6
C***TYPE      DOUBLE PRECISION (BSPVD-S, DBSPVD-D)
C***KEYWORDS  DIFFERENTIATION OF B-SPLINE, EVALUATION OF B-SPLINE
C***AUTHOR  Amos, D. E., (SNLA)
C***DESCRIPTION
C
C     Written by Carl de Boor and modified by D. E. Amos
C
C     Abstract    **** a double precision routine ****
C
C         DBSPVD is the BSPLVD routine of the reference.
C
C         DBSPVD calculates the value and all derivatives of order
C         less than NDERIV of all basis functions which do not
C         (possibly) vanish at X.  ILEFT is input such that
C         T(ILEFT) .LE. X .LT. T(ILEFT+1).  A call to INTRV(T,N+1,X,
C         ILO,ILEFT,MFLAG) will produce the proper ILEFT.  The output of
C         DBSPVD is a matrix VNIKX(I,J) of dimension at least (K,NDERIV)
C         whose columns contain the K nonzero basis functions and
C         their NDERIV-1 right derivatives at X, I=1,K, J=1,NDERIV.
C         These basis functions have indices ILEFT-K+I, I=1,K,
C         K .LE. ILEFT .LE. N.  The nonzero part of the I-th basis
C         function lies in (T(I),T(I+K)), I=1,N).
C
C         If X=T(ILEFT+1) then VNIKX contains left limiting values
C         (left derivatives) at T(ILEFT+1).  In particular, ILEFT = N
C         produces left limiting values at the right end point
C         X=T(N+1).  To obtain left limiting values at T(I), I=K+1,N+1,
C         set X= next lower distinct knot, call INTRV to get ILEFT,
C         set X=T(I), and then call DBSPVD.
C
C     Description of Arguments
C         Input      T,X are double precision
C          T       - knot vector of length N+K, where
C                    N = number of B-spline basis functions
C                    N = sum of knot multiplicities-K
C          K       - order of the B-spline, K .GE. 1
C          NDERIV  - number of derivatives = NDERIV-1,
C                    1 .LE. NDERIV .LE. K
C          X       - argument of basis functions,
C                    T(K) .LE. X .LE. T(N+1)
C          ILEFT   - largest integer such that
C                    T(ILEFT) .LE. X .LT.  T(ILEFT+1)
C          LDVNIK  - leading dimension of matrix VNIKX
C
C         Output     VNIKX,WORK are double precision
C          VNIKX   - matrix of dimension at least (K,NDERIV) contain-
C                    ing the nonzero basis functions at X and their
C                    derivatives columnwise.
C          WORK    - a work vector of length (K+1)*(K+2)/2
C
C     Error Conditions
C         Improper input is a fatal error
C
C***REFERENCES  Carl de Boor, Package for calculating with B-splines,
C                 SIAM Journal on Numerical Analysis 14, 3 (June 1977),
C                 pp. 441-472.
C***ROUTINES CALLED  DBSPVN, XERMSG
C***REVISION HISTORY  (YYMMDD)
C   800901  DATE WRITTEN
C   890531  Changed all specific intrinsics to generic.  (WRB)
C   890831  Modified array declarations.  (WRB)
C   890831  REVISION DATE from Version 3.2
C   891214  Prologue converted to Version 4.0 format.  (BAB)
C   900315  CALLs to XERROR changed to CALLs to XERMSG.  (THJ)
C   920501  Reformatted the REFERENCES section.  (WRB)
C***END PROLOGUE  DBSPVD