CERN Accelerating science

This website is no longer maintained. Its content may be obsolete. Please visit http://home.cern/ for current CERN information.



next up previous
Next: Obsolete Subprograms Up: Program Library News Previous: Reissued Subprograms

New/Changed Subprograms

  F. James ( V115 -- V122) and K.S. Kölbig CN/ASD ( E408 -- H301)

E408 RCHPWS: Conversion of Chebyshev to Power and Power to Chebyshev Series
A new subroutine subprogram E408 RCHPWS (Conversion of Chebyshev to power and power to Chebyshev series) with entries RCHPWS and RPWCHS, written in Fortran, which performs the conversion of a finite Chebyshev series to a finite power series (a polynomial) and vice versa, has been submitted to MATHLIB for CDC and Cray computers. A double-precision version DCHPWS with entries DCHWPS and DPWCHS is provided on IBM and similar computers. This subprogram replaces subroutine subprogram E401 ECTRAD in MATHLIB, which thus becomes obsolete. The old routine will be left in MATHLIB for a few months and will then be deleted.
H101 RSMPLX: Linear Optimization Using the Simplex Algorithm
A new subroutine subprogram H101 RSMPLX (Linear optimization using the simplex algorithm) written in Fortran, which calculates a vector for which a certain linear form assumes a maximum under linear inequality or equality constraints, has been submitted to MATHLIB for CDC and Cray computers. A double-precision version DSMPLX is provided on IBM and similar computers. This subprogram replaces subroutine subprogram H100 SIMPLE in MATHLIB, which thus becomes obsolete. The old routine will be left in MATHLIB for a few months and will then be deleted.
H301 ASSNDX: Assignment Problem
A new subroutine subprogram H301 ASSNDX (Assignment problem) written in Fortran, which selects certain elements of a rectangular matrix such that the sum of them is minimal, has been submitted to MATHLIB. This subprogramm replaces subroutine subprogram H300 ASSIGN in MATHLIB, which thus becomes obsolete. The old routine will be left in MATHLIB for a few months and will then be deleted.
V115 RANLUX: Uniform Random Numbers of Guaranteed Quality
A new subroutine subprogram V115 RANLUX (Uniform random numbers of garanteed quality) written in Fortran, has been submitted to KERNLIB. It generates pseudorandom numbers uniformly distributed in the interval (0,1), the end points excluded. Each call produces an array of single-precision real numbers of which 24 bits of mantissa are random. The user can choose a luxury level which guarantees the quality required for his application. The lowest luxury level (zero) gives a fast generator which will fail some sophisticated tests of randomness; the highest level (four) is about five times slower but guarantees complete randomness. In all cases the period is greater than . Independent subsequences can be generated. Entries are provided for initialization and checkpointing.
V116 RM48: Double-Precision Uniform Random Numbers
A new subroutine subprogram V116 RM48 (Double-Precision Uniform random numbers) written in Fortran, has been submitted to KERNLIB. It generates pseudorandom numbers using a double-precision (64-bit) adaptation of V113 RANMAR. The floating-point numbers in the interval (0,1), the end points excluded, have 48 significant bits of mantissa (additional bits of mantissa, if supported by the hardware, are zero). Both the code and the results are portable.
V120 RNORML: Gaussian-distributed Random Numbers
New subroutine subprogram V120 RNORML and RNORMX (Gaussian-distributed random numbers) written in Fortran, has been submitted to KERNLIB. RNORML and RNORMX generate (vectors of) single-precision random numbers in a Gaussian distribution of mean zero and variance one. RNORML uses the uniform generator V113 RANMAR underneath, and RNORMX allows the user to choose the uniform generator to be used underneath. The code is portable, but the results are not guaranteed to be identical on all platforms because there is branch on a floating-point compare which may (very rarely) cause the sequence produced on a given platform to be out of step with that of a different platform.
V122 CORSET: Correlated Gaussian-distributed Random Numbers
New subroutine subprogram V122 CORSET and CORGEN (Correlated Gaussian-distributed random numbers) written in Fortran, has been submitted to MATHLIB. The subprogram CORGEN generates vectors of single-precision random numbers in a Gaussian distribution of mean zero and covariance matrix V. The generator must first be set up by a call to CORSET which transforms the covariance matrix V to an appropriate square root matrix C which is then used by CORGEN. CORGEN uses the Gaussian generator V120 RNORML underneath, which in turn uses the uniform generator V113 RANMAR underneath. The code is portable, but the results are not guaranteed to be identical on all platforms as explained in V120 RNORML.
Subprograms V115 RANLUX, V116 RM48, V120 RNORML and V122 CORSET replace older subprograms for the generation of random numbers, which thus become obsolete. For details see Section 6.12.



next up previous
Next: Obsolete Subprograms Up: Program Library News Previous: Reissued Subprograms



Janne Saarela
Tue May 16 14:52:44 METDST 1995