Note:  In the following, when a filename is of the form "aaaaX.f", the
       letter "X" stands for either an "S", "D", or a "C", depending on
       whether you are obtaining the single precision, double precision, 
       or complex arithmetic version of the CgCode.

------------------------------------------------------------------------------

This directory contains a set of routines to:

(1)  Discretize a standard PDE.
(2)  Set up the corresponding system Ax=B to solve.
(3)  Call CgCode to solve the linear system.
(4)  Print statistics and numerical solution (if desired).

The files are:

(0)  cgX.tar.gz        - The original gzipped-tar file you downloaded.
(1)  README            - A guide to the distributed code (this file).
(2)  INTRODUCTION      - An introduction to CgCode.
(3)  makefile          - Simple Makefile 
(4)  cgcodeX.f         - The CgCode software package.
(5)  pdeX.f            - Driver test program, with primary setup routines.
(6)  cgnumlibX.f       - Necessary BLAS/LINPACK/EISPACK routines (see note 2).
(7)  in                - Input parameters.
(8)  out               - Output run statistics from driver program.
(9)  outn              - Numerical solution plus CgCode output.

------------------------------------------------------------------------------

Note 1: To obtain and use CgCode and the test program, either transfer
        each of the above files via ftp, or simply transfer the single  
        file "cgX.shar", and on your UNIX system extract the other files
        by entering the command:   "/bin/shar cgX.shar".

        If you have received the code by electronic mail, you were sent
        only the shar file "cgX.shar" containing all the other files.

Note 2: The distribution you just received contains only a comment block
        in the file called "cgnumlibX.f", the file that should contain the
        standard numerical library routines that CgCode needs, including
        various LINPACK, EISPACK, and BLAS subroutines.  

        CgCode is very sensitive to the correct and efficient 
        performance of these routines, and it is extremely important
        that these routines be correctly installed on your computer.  
        Therefore, we do not usually provide these routines.  See your 
        system administrator for more information on how to access 
        them if they have been installed already (and they probably
        have been if there are other numerical software packages 
        running at your site), or send electronic mail to netlib to 
        obtain the software if it is not available on your system.

        If you still have trouble obtaining these routines, we may 
        be able to provide some of them.  However, as we may not be 
        familiar with your particular architecture, the performance 
        of CgCode may suffer if the routines are not properly installed.

