Lines Matching +full:single +full:- +full:shot

9  *   dividend -- the thing being divided
10  *   divisor -- how many ways to divide it
12  *   N -- how many bits per iteration we try to get
14  *   WORDSIZE -- how many bits altogether we're talking about:
17  *   TOPBITS -- how many bits are in the top "decade" of a number:
18  *        define(TOPBITS, eval( WORDSIZE - N*((WORDSIZE-1)/N) ) )
20  *   Q -- the partial quotient under development -- initially 0
21  *   R -- the remainder so far -- initially == the dividend
22  *   ITER -- number of iterations of the main division loop which will
25  *   V -- the current comparand -- initially divisor*2ˆ(ITER*N-1)
27  *   current estimate for non-large dividend is
29  *   a large dividend is one greater than 2ˆ(31-TOPBITS) and takes a
47  *   $1 -- the current depth, 1<=$1<=N
48  *   $2 -- the current accumulation of quotient bits
49  *   N -- max depth
52  *   R -- current remainder
53  *   Q -- current quotient
54  *   V -- current comparand
55  *   cc -- set on current value of R
78 		add	Q, ($2*2-1), Q
79 	',` DEVELOP_QUOTIENT_BITS( incr($1), `eval(2*$2-1)')
135 	sethi	%hi(1<<(WORDSIZE-TOPBITS-1)),T
140 	! Here, the dividend is >= 2ˆ(31-N) or so. We must be careful here,
141 	! as our usual N-at-a-shot divide step will cause overflow and havoc.
158 		! This means that R has the high-order bit set.
174 ! do single-bit divide steps
178 ! and are only done if R >= 0. Because both R and V may have the high-
229 	! non-restoring fixup if remainder < 0, otherwise annulled
240 `	neg	%o2,%o2			! Q <- -Q
241 ',`	neg	%o3,%o3 		! R <- -R
244 	retl				! leaf-routine return
246 `	mov	%o2,%o0			! quotient <- Q
247 ',`	mov	%o3,%o0			! remainder <- R