Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site ncoast.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!qantel!lll-crg!ucdavis!ucbvax!decvax!cwruecmp!hal!ncoast!simpsong
From: simpsong@ncoast.UUCP (Gregory R. Simpson @ The North Coast)
Newsgroups: net.sources
Subject: matrix mult. (source and question)
Message-ID: <890@ncoast.UUCP>
Date: Thu, 14-Nov-85 00:18:16 EST
Article-I.D.: ncoast.890
Posted: Thu Nov 14 00:18:16 1985
Date-Received: Mon, 18-Nov-85 06:01:53 EST
Distribution: net
Organization: Cleveland Public Access UNIX, Cleveland, OH
Lines: 120

*** Line eater poison ***

Below is a kludge version of a matrix multiplier in C. The primary
reason for posting this is in hope that someone out there knows the
proper method of doing it. The problem I encountered was that in order
to pass a two dimensional array (hence a matrix) I needed to specify
the second bound of the array. However, I want this to be a generic
matrix multipler, for matrices of all sizes not just a fixed size. Since
the column increases most rapidly in a two dimensional array, declaring
a large array (like mat[][20]) puts all of a 3x3 into the first row.
In the much dreaded language of Fortran, one may simply pass a variable
to the dimension statement... this doesn't work in C. I believe this
is a job for pointers, but I was unsuccessful in getting them to work.
Please give me pointers on pointers (pun intended) for this application.
(or if you are ambitious send me a better version of code)

My fixes were fairly, crude, and the result is returned in a mxn vector
as opposed to an mxn matrix. Please send any pointers, references, code
help, flames to the address below. I have included my version below.

p.s. Once this issue is resolved, I will post an enhanced matrix
manipulation set, (mult., inversion, determinant, co-factors, etc.)

Thanks in advance, 

      Gregory R. Simpson

      UUCP: decvax!cwruecmp!ncoast!simpsong
      CSNET: simpsong@case.CSNET
      ARPA: simpsong@purdue-ecn.ARPA

p.s. I am also looking for the manual page to roff.c posted from site
gloria about 2 months back...

======= bend - mutilate - cut - line-eat - here - ok - ================

/* a sample main program to call matmult */

main()
{
       float       a[3][3],b[3][3];
       float       c[9];
       int         i,j;

/* a typical array */

       a[0][0]=1.0;
       a[0][1]=2.0;
       a[0][2]=3.0;
       a[1][0]=4.0;
       a[1][1]=5.0;
       a[1][2]=6.0;
       a[2][0]=7.0;
       a[2][1]=8.0;
       a[2][2]=9.0;
       
/* another typical array */

       for (i=0;i<=2;i++)       {
         for (j=0;j<=2;j++)       {
           b[i][j] = a[i][j];
         }
       }

/* multiply two typical arrays together */

       matmlt(a,b,c,3,3,3);

/* print the result, (it is in a 3 times 3 vector... boo hiss) */

       for (i=0;i<=8;i++)       {
               printf("%f\n",c[i]);
       }
}

/* the matrix multiplier */

matmlt(a,b,c,n,m,l)

/* this is a lot easier in fortran, cause then I can say dimension a(m,n) */

int       n, m, l;
float     a[], b[];
float     c[];

{
       float     a2[10][10], b2[10][10], c2[10][10];
       int       i,j,k,inc;

/* load the arrays I passed into an array - kludgy ehhh? */

       inc = 0;
       for (i=0;i<=2;i++)       {
         for (j=0;j<=2;j++)       {
           a2[i][j] = a[inc];
           b2[i][j] = b[inc];
           inc++;
         }
       }                              
       
       for (i = 0; i <= n-1 ; i++) {
         for (k = 0; k <= l-1 ; k++) {
            c2[i][k] = 0;
           for (j = 0; j <= m-1 ; j++) {
             c2[i][k] = (a2[i][j])*(b2[j][k]) + c2[i][k];
           }
         }
       }

/* load the array result back into a vector to pass back */

       inc = 0;
       for (i=0;i<=2;i++)       {
         for (j=0;j<=2;j++)       {
           c[inc] = c2[i][j];
           printf("%f  %f  %f \n",a2[i][j],b2[i][j],c2[i][j]);
           inc++;
         }
       }  
}