/*
	This C language subroutine computes the 
	product of two matricies.

		     s
	z(i,j) = SUMMATION[x(i,k)*y(k,j)]
		    k=1

	with i= 1 to r, and j=1 to t

	
	The Run Time Library for the C Language.

	Gordon A. Sterling	(617) 461 - 3076
	DSP Development Tools Engineering

	Created on 5/25/90	
	Updated on 5/94 by AS

	#include <matrix.h>							    <Header>
	float matmul(float z[][], float x[][], float y[][], int r, int s ,int t);   <Prototype>
*/

#include "lib_glob.h"
#include "mat_glob.h"

.SEGMENT/CODE	Code_Space_Name;
.FILE 		RTL_FILENAME;

.GLOBAL	    	___matmulDPD;

___matmulDPD:	
matmul_core:	R2=dma_i;
		put(R2);
		R2=dmb_i;
		put(R2);
		R2=pma_i;
		put(R2);
		R2=pmb_i;
		put(R2);
		put(R5);
		put(R6);
		put(R7);
		put(R13);
		put(R14);
		put(R15);

	    	readparam4(R6);
		readparam5(R5);
	    	readparam6(R7);

		R13=PASS R12, R12=R4;	
		R14=PASS R8, R2=R4;

dmpmdm_mul:	pm_lnt=R5;
		dm_mdf=R7;
		R15=PASS R13, dmb_i=R12;	/* Point to Z and start of Y	*/
		LCNTR=R6, DO dmpmdm_row UNTIL LCE;/* Loop for each outrow	*/
		    pm_bse=R14;			/* Point to X row		*/
		    LCNTR=R7, DO dmpmdm_column UNTIL LCE;/* Loop for each outcol*/
		  	R12=R12-R12, dma_i=R15;	/* Point to place in Y	*/
		    	R8=R8-R8, F0=PM(pm_ptr, 1);	/* Read first X		*/
			R6=R5-1, F4=DM(dma_i, dm_mdf);	/* Read first Y		*/
			LCNTR=R6, DO dmpmdm_element UNTIL LCE;
			    F8=F0*F4, F12=F8+F12, F0=PM(pm_ptr, pm_1);
dmpmdm_element:		    F4=DM(dma_i, dm_mdf);
			F8=F0*F4, F12=F8+F12;		/* Compute final mult	*/
			F12=F8+F12;			/* Compute final sum	*/
dmpmdm_column:		R15=R15+1, DM(dmb_i,1)=F12;	/* Advance Y ptr	*/
dmpmdm_row:    	    R14=R14+R5, R15=R13;	/* Point to next X, start of y	*/

		R0=PASS R2, pm_lnt=pm_0;	/* Set return value		*/

restore_state:	FETCH_RETURN
		get(R15,1);
		get(R14,2);
		get(R13,3);
		get(R7,4);
		get(R6,5);
		get(R5,6);
		get(pmb_i,7);
		get(pma_i,8);
		get(dmb_i,9);
		get(dma_i,10);
		RETURN (DB);
		RESTORE_STACK
		RESTORE_FRAME
.ENDSEG;