B002 PRMFCT: Prime Numbers and Prime Factor Decomposition
A new function subprogram B002 PRMFCT (Prime numbers and prime
factor decomposition) written in Fortran, which (i) sets the first
prime numbers into an array, (ii) performes a decomposition
of a positive integer N<10007 into its prime factors, or (iii)
performes a decomposition of the factorial
() into its
prime factors, has been submitted to MATHLIB.
C317 CPSIPG: Psi (Digamma) and Polygamma Functions for
Complex Argument
A new function subprogram C317 CPSIPG (Psi (digamma) and
polygamma functions for complex argument)
written in Fortran, which calculates the psi or digamma function
, where
is the gamma function, and the polygamma functions
for and complex values ,
has been submitted to MATHLIB.
A COMPLEX*16 version WPSIPG
is provided on machines which support this data type.
This function subprogram replaces C307 CDIGAM with entries
CDIGAM and WDIGAM in MATHLIB, which thus becomes obsolete.
The old routine will be left in MATHLIB for a few months and will then
be deleted.
D120 RADMUL: Adaptive Quadrature for Multiple Integrals
over -Dimensional Rectangular Regions
A new subroutine subprogram D120 RADMUL (Adaptive quadrature
for multiple integrals over -dimensional rectangular regions)
written in Fortran, which calculates, to an attempted
specified accuracy, the value of a more-dimensional integral
by using certain five- and seven-point rules together with a
subdivision strategy, has been submitted to MATHLIB for Cray and
similar computers. A double-precision version DADMUL is
provided on IBM and similar computers.
E106 LOCATR:Binary Search for Element in Ordered Array
An entry LOCATD, which performs a binary search in an array of
non-decreasing double precision numbers has been added to function
subprogram E106 LOCATR (Binary search for element in ordered
array) in KERNLIB for IBM and similar computers. Further, the entry
LOCATF has been changed to LOCATR for consistency.
The old entry will be left in KERNLIB for a certain (sufficiently
long) period and will then be deleted.
F003 RMADD:Elementary Matrix Processing
Because of a name clash with an entry in F001 LAPACK,
the User Entry Names RRSCL, DRSCL and CRSCL in the
KERNLIB subprogram package F003 RMADD (Elementary Matrix
Processing), which perform the multiplication
of the elements
of a matrix by the elements
of a vector,
have been changed to RMDMP, DMDMP, and CMDMP,
respectively. In their own interest, users are invited to adapt their
programs without delay.