Public access code is software and documentation that is made available for non-commercial use but in which NAG, or the original author, retains the Copyright.
A new Graphical User Interface to some NAGWare f77 and f90 Tools is now available free. The GUI is intended to guide the user through some useful tools for converting fixed format Fortran 77 to Fortran 90. Note that both NAGWare f77 and f90 Tools are required to fully utilise the GUI.
The DaPre_77 system is a pre-processor for FORTRAN 77 code that produces, from a user's FORTRAN 77 function, FORTRAN subroutines for calculating derivatives.
The software was developed by Dr John Pryce and Bruce Stephens of the Royal Military College of Science and NAG Ltd. The software is provided for free distribution over Internet at the NAG WWW/Gopher server, no technical support is provided.
This software is provided in a form suitable for Sun 4 (SunOS) machine, please see the Installer's Note (IN) and Users' Note (UN) for more details.
Q387 is not a NAG product. It is "protected shareware" that emulates a maths coprocessor for 80386 or 80486SX based PCs. It will enable release 4 of GLIM and release 3.1 of Genstat 5 to run on PCs without 80387 hardware capabilites. Q387 is Copyright, 1991-1994, QuickWare, Austin, Texas. (See Readme or documentation with product).
Q387 Version 3.66 is a math accelerator and math coprocessor emulator for computers which do not have a coprocessor, such as 486SX and 486SLC computers. It will both increase the performance of your computer with most math intensive applications and allow you to run applications which require a math coprocessor. Q387 requires a 386SX or higher processor, at least 1.5 Megabytes total memory, and DOS 5 or 6, or QEMM386. Q387 is compatible with DOS 5 and 6, QEMM386, DOS Extender applications, and Windows 3.0, 3.1, and Windows for Workgroups 3.11.
The f2c output polishing toolset was described in an article by George Levy in the February 1995 issue of Software Practice and Experience. The software is available on-line.
Note: This document does not give information about the public domain software Toolpack/1 as it has been superseded by the NAGWare f77 Tools (and the yet to be released NAGWare f90 Tools). However, as the NAGWare products are not yet available on all machine ranges, the Toolpack/1 software may still be required. Please contact NAG for further information.
Note that NAG distributes only the software and any machine-readable documents described below; the User Guides are obtainable from their publishers.
LAPACK is a public-domain linear algebra package intended to supersede Linpack and Eispack and run efficiently on a wide range of modern high-performance computers.
It is a collection of transportable Fortran 77 subroutines for solving the most common computational problems in linear algebra, such as systems of linear equations, linear least-squares problems, eigenvalue problems and singular value problems. It can also handle many associated computations such as matrix factorizations or estimating condition numbers. Dense and banded matrices are provided for, but not general sparse matrices. In all areas similar functionality is provided for real and complex matrices.
LAPACK is designed to be efficient on vector processors, high-performance "superscalar" workstations, and shared-memory multiprocessors. LAPACK can also be used satisfactorily on all types of scalar machines (PCs, workstations, mainframes).
Efficiency is achieved by constructing LAPACK routines (as far as possible) from calls to the BLAS (Basic Linear Algebra Subprograms), especially the Level 3 BLAS for matrix-matrix operations and the Level 2 BLAS for matrix-vector operations. Very efficient implementations of the BLAS are already provided by many of the major vendors of high-performance computers, and thus the BLAS act as a portability-base for LAPACK, providing efficiency without compromising the portability of the LAPACK software.
LAPACK has been developed by a collaborative project, involving several scientists in the USA (funded primarily by the National Science Foundation) and NAG Ltd. Many other people have contributed ideas or software, or have tested preliminary versions of the package.
Individual routines from LAPACK are most easily obtained by electronic mail through netlib. The complete package, including test code and timing programs, constitutes some 600,000 lines of Fortran, and is being distributed on magnetic media by NAG, on a cost-recovery basis. A comprehensive Installation Guide is included on the release media. The documentation for the package is contained in the following book:
[1] Anderson, E., Bai, Z., Bischof, C., Demmel, J., Dongarra, J., Du Croz,
J.J., Greenbaum, A., Hammarling, S., McKenney, A., Ostrouchov, S. and
Sorensen, D.
LAPACK Users' Guide
SIAM, Philadelphia, 1992
This is a collection of Fortran subroutines which compute the eigenvalues and eigenvectors of nine classes of matrices. The package can determine the eigensystem of complex general, Hermitian, real general, real symmetric, real symmetric band, real symmetric tridiagonal, special real tridiagonal, generalized real, and generalized real symmetric matrices. In addition, there are two routines to compute the singular value decomposition and to solve certain least-squares problems.
The routines are supplied on magnetic tape in single and double precision, with test programs, data and installation instructions.
The documentation for the codes is contained in the following books:
[1] Smith, B.T., Boyle, J.M., Dongarra, J.J., Garbow, B.S., Ikebe, Y., Klema,
V.C. and Moler, C.B.
Matrix Eigensystem Routines - EISPACK Guide, Second Edition
Lecture Notes in Computer Science, Volume 6, Springer-Verlag, 1976.
[2] Garbow, B.S., Boyle, J.M., Dongarra, J.J., and Moler, C.B.
Matrix Eigensystem Routines - EISPACK Guide Extension
Lecture Notes in Computer Science, Volume 51, Springer-Verlag, 1977.
This is a collection of Fortran subroutines which analyse and solve various linear equations and linear least-squares problems. The package solves linear systems where matrices are general square, banded, symmetric indefinite, symmetric positive-definite, triangular or tridiagonal. In addition the package computes the QR and singular value decompositions of rectangular matrices and applies them to linear least-squares problems. Single precision, double precision and complex versions of the code are included. The tape distributed contains the Fortran source for LINPACK, the Basic Linear Algebra Subprograms (BLAS) needed by LINPACK, testing aids and program comments.
The complete documentation for the package can be found in:
[1] Dongarra, J.J., Bunch, J.R., Moler, C.B. and Stewart, G.W.
LINPACK Users' Guide,
SIAM, Philadelphia, 1979.
This is a collection of Fortran subroutines for the numerical solution of systems of nonlinear equations and nonlinear least-squares problems. For each problem area MINPACK contains algorithms that proceed either from an analytical specification of the Jacobian matrix of the problem functions or directly from the problem functions themselves. Since the specification of the Jacobian matrix can be an error-prone task, MINPACK contains an algorithm to check that the Jacobian matrix is consistent with the function.
The MINPACK tape includes machine-readable documentation and a set of testing aids.
[ Home : InfoDesk : GuestBook ] Last modified: Tue Mar 5 17:00:00 1996
Copyright The Numerical Algorithms Group Ltd, Oxford UK. 1995 <webmaster@nag.co.uk>