testNumHi	equ	$1CE		; Q value high word
testNumLo	equ	$4000		; Q value low word
QL		equ	17		; stack offset for Q lower word
QH		equ	15		; stack offset for Q upper word
SAR		equ	7		; stack offset for successive approximation register
mask		equ	5		; stack offset for mask for performing approximation
buf_lo	equ	3		; stack offset for buffer lower word
buf_hi	equ	1		; stack offset for buffer upper word
lp_cnt	equ	0		; stack offset for loop count
locVAR	equ	9		; space for local variables

		org	$1000
sq_root	ds.w	1
		org	$1500
		lds	#$1500
		ldd	#testNumLo
		pshd
		ldd	#testNumHi
		pshd
		jsr	SqRoot32b	; compute the square root of a 32-bit number
		leas	4,SP		; deallcate space used by parameters
		std	sq_root	; save the square root
		swi
; ****************************************************************************
; The following subroutine computes the square root of a 32-bit integer using
; the sucessive approximation method. The 32-bit number is pushed to this 
; subroutine.
; ****************************************************************************
SqRoot32b	pshx
		pshy
		leas	-locVAR,SP	; allocate local variables
		movb	#16,lp_cnt,sp	; initialize loop count
		movw	#$8000,mask,SP	
		movw	#0,SAR,SP	; initialize SAR to 0
sqLoop	ldd	mask,SP
		anda	SAR,SP	; guess the current bit to be 1
		andb	SAR+1,SP	;	"	
		tfr	D,Y
		emul			; compute the square root
		subd	QL,SP		; compute the SAR*SAR with Q
		tfr	Y,D		;	"
		sbcb	QH+1,SP	;	"
		sbca	QH,SP		;	"
		bcs	guessR	; guess correctly for bit i
		bra	nextbit
guessR	ldd	mask,SP
		andb	SAR+1,SP	; set bit i to 1
		anda	SAR,SP	;	"
		std	SAR,SP
nextbit	ldd	mask,SP	; set the bit i - 1 of mask to 1
		lsrd			;	"
		std	mask,SP	;	"
		dec	lp_cnt,SP	; have we done with all 16 bits?
		bne	sqLoop
		ldd	SAR,SP
		leas	locVAR,SP
		puly
		pulx
		rts
		end