xref: /openbsd/sys/arch/sparc64/fpu/fpu_sqrt.c (revision 097a140d)
1 /*	$OpenBSD: fpu_sqrt.c,v 1.5 2021/03/11 11:17:00 jsg Exp $	*/
2 /*	$NetBSD: fpu_sqrt.c,v 1.2 1994/11/20 20:52:46 deraadt Exp $ */
3 
4 /*
5  * Copyright (c) 1992, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This software was developed by the Computer Systems Engineering group
9  * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
10  * contributed to Berkeley.
11  *
12  * All advertising materials mentioning features or use of this software
13  * must display the following acknowledgement:
14  *	This product includes software developed by the University of
15  *	California, Lawrence Berkeley Laboratory.
16  *
17  * Redistribution and use in source and binary forms, with or without
18  * modification, are permitted provided that the following conditions
19  * are met:
20  * 1. Redistributions of source code must retain the above copyright
21  *    notice, this list of conditions and the following disclaimer.
22  * 2. Redistributions in binary form must reproduce the above copyright
23  *    notice, this list of conditions and the following disclaimer in the
24  *    documentation and/or other materials provided with the distribution.
25  * 3. Neither the name of the University nor the names of its contributors
26  *    may be used to endorse or promote products derived from this software
27  *    without specific prior written permission.
28  *
29  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
30  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
31  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
32  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
33  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
34  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
35  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
36  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
37  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
38  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
39  * SUCH DAMAGE.
40  *
41  *	@(#)fpu_sqrt.c	8.1 (Berkeley) 6/11/93
42  */
43 
44 /*
45  * Perform an FPU square root (return sqrt(x)).
46  */
47 
48 #include <sys/types.h>
49 
50 #include <machine/reg.h>
51 
52 #include <sparc64/fpu/fpu_arith.h>
53 #include <sparc64/fpu/fpu_emu.h>
54 
55 /*
56  * Our task is to calculate the square root of a floating point number x0.
57  * This number x normally has the form:
58  *
59  *		    exp
60  *	x = mant * 2		(where 1 <= mant < 2 and exp is an integer)
61  *
62  * This can be left as it stands, or the mantissa can be doubled and the
63  * exponent decremented:
64  *
65  *			  exp-1
66  *	x = (2 * mant) * 2	(where 2 <= 2 * mant < 4)
67  *
68  * If the exponent `exp' is even, the square root of the number is best
69  * handled using the first form, and is by definition equal to:
70  *
71  *				exp/2
72  *	sqrt(x) = sqrt(mant) * 2
73  *
74  * If exp is odd, on the other hand, it is convenient to use the second
75  * form, giving:
76  *
77  *				    (exp-1)/2
78  *	sqrt(x) = sqrt(2 * mant) * 2
79  *
80  * In the first case, we have
81  *
82  *	1 <= mant < 2
83  *
84  * and therefore
85  *
86  *	sqrt(1) <= sqrt(mant) < sqrt(2)
87  *
88  * while in the second case we have
89  *
90  *	2 <= 2*mant < 4
91  *
92  * and therefore
93  *
94  *	sqrt(2) <= sqrt(2*mant) < sqrt(4)
95  *
96  * so that in any case, we are sure that
97  *
98  *	sqrt(1) <= sqrt(n * mant) < sqrt(4),	n = 1 or 2
99  *
100  * or
101  *
102  *	1 <= sqrt(n * mant) < 2,		n = 1 or 2.
103  *
104  * This root is therefore a properly formed mantissa for a floating
105  * point number.  The exponent of sqrt(x) is either exp/2 or (exp-1)/2
106  * as above.  This leaves us with the problem of finding the square root
107  * of a fixed-point number in the range [1..4).
108  *
109  * Though it may not be instantly obvious, the following square root
110  * algorithm works for any integer x of an even number of bits, provided
111  * that no overflows occur:
112  *
113  *	let q = 0
114  *	for k = NBITS-1 to 0 step -1 do -- for each digit in the answer...
115  *		x *= 2			-- multiply by radix, for next digit
116  *		if x >= 2q + 2^k then	-- if adding 2^k does not
117  *			x -= 2q + 2^k	-- exceed the correct root,
118  *			q += 2^k	-- add 2^k and adjust x
119  *		fi
120  *	done
121  *	sqrt = q / 2^(NBITS/2)		-- (and any remainder is in x)
122  *
123  * If NBITS is odd (so that k is initially even), we can just add another
124  * zero bit at the top of x.  Doing so means that q is not going to acquire
125  * a 1 bit in the first trip around the loop (since x0 < 2^NBITS).  If the
126  * final value in x is not needed, or can be off by a factor of 2, this is
127  * equivalent to moving the `x *= 2' step to the bottom of the loop:
128  *
129  *	for k = NBITS-1 to 0 step -1 do if ... fi; x *= 2; done
130  *
131  * and the result q will then be sqrt(x0) * 2^floor(NBITS / 2).
132  * (Since the algorithm is destructive on x, we will call x's initial
133  * value, for which q is some power of two times its square root, x0.)
134  *
135  * If we insert a loop invariant y = 2q, we can then rewrite this using
136  * C notation as:
137  *
138  *	q = y = 0; x = x0;
139  *	for (k = NBITS; --k >= 0;) {
140  * #if (NBITS is even)
141  *		x *= 2;
142  * #endif
143  *		t = y + (1 << k);
144  *		if (x >= t) {
145  *			x -= t;
146  *			q += 1 << k;
147  *			y += 1 << (k + 1);
148  *		}
149  * #if (NBITS is odd)
150  *		x *= 2;
151  * #endif
152  *	}
153  *
154  * If x0 is fixed point, rather than an integer, we can simply alter the
155  * scale factor between q and sqrt(x0).  As it happens, we can easily arrange
156  * for the scale factor to be 2**0 or 1, so that sqrt(x0) == q.
157  *
158  * In our case, however, x0 (and therefore x, y, q, and t) are multiword
159  * integers, which adds some complication.  But note that q is built one
160  * bit at a time, from the top down, and is not used itself in the loop
161  * (we use 2q as held in y instead).  This means we can build our answer
162  * in an integer, one word at a time, which saves a bit of work.  Also,
163  * since 1 << k is always a `new' bit in q, 1 << k and 1 << (k+1) are
164  * `new' bits in y and we can set them with an `or' operation rather than
165  * a full-blown multiword add.
166  *
167  * We are almost done, except for one snag.  We must prove that none of our
168  * intermediate calculations can overflow.  We know that x0 is in [1..4)
169  * and therefore the square root in q will be in [1..2), but what about x,
170  * y, and t?
171  *
172  * We know that y = 2q at the beginning of each loop.  (The relation only
173  * fails temporarily while y and q are being updated.)  Since q < 2, y < 4.
174  * The sum in t can, in our case, be as much as y+(1<<1) = y+2 < 6, and.
175  * Furthermore, we can prove with a bit of work that x never exceeds y by
176  * more than 2, so that even after doubling, 0 <= x < 8.  (This is left as
177  * an exercise to the reader, mostly because I have become tired of working
178  * on this comment.)
179  *
180  * If our floating point mantissas (which are of the form 1.frac) occupy
181  * B+1 bits, our largest intermediary needs at most B+3 bits, or two extra.
182  * In fact, we want even one more bit (for a carry, to avoid compares), or
183  * three extra.  There is a comment in fpu_emu.h reminding maintainers of
184  * this, so we have some justification in assuming it.
185  */
186 struct fpn *
187 fpu_sqrt(fe)
188 	struct fpemu *fe;
189 {
190 	register struct fpn *x = &fe->fe_f1;
191 	register u_int bit, q, tt;
192 	register u_int x0, x1, x2, x3;
193 	register u_int y0, y1, y2, y3;
194 	register u_int d0, d1, d2, d3;
195 	register int e;
196 
197 	/*
198 	 * Take care of special cases first.  In order:
199 	 *
200 	 *	sqrt(NaN) = NaN
201 	 *	sqrt(+0) = +0
202 	 *	sqrt(-0) = -0
203 	 *	sqrt(x < 0) = NaN	(including sqrt(-Inf))
204 	 *	sqrt(+Inf) = +Inf
205 	 *
206 	 * Then all that remains are numbers with mantissas in [1..2).
207 	 */
208 	if (ISNAN(x) || ISZERO(x))
209 		return (x);
210 	if (x->fp_sign)
211 		return (fpu_newnan(fe));
212 	if (ISINF(x))
213 		return (x);
214 
215 	/*
216 	 * Calculate result exponent.  As noted above, this may involve
217 	 * doubling the mantissa.  We will also need to double x each
218 	 * time around the loop, so we define a macro for this here, and
219 	 * we break out the multiword mantissa.
220 	 */
221 #ifdef FPU_SHL1_BY_ADD
222 #define	DOUBLE_X { \
223 	FPU_ADDS(x3, x3, x3); FPU_ADDCS(x2, x2, x2); \
224 	FPU_ADDCS(x1, x1, x1); FPU_ADDC(x0, x0, x0); \
225 }
226 #else
227 #define	DOUBLE_X { \
228 	x0 = (x0 << 1) | (x1 >> 31); x1 = (x1 << 1) | (x2 >> 31); \
229 	x2 = (x2 << 1) | (x3 >> 31); x3 <<= 1; \
230 }
231 #endif
232 #if (FP_NMANT & 1) != 0
233 # define ODD_DOUBLE	DOUBLE_X
234 # define EVEN_DOUBLE	/* nothing */
235 #else
236 # define ODD_DOUBLE	/* nothing */
237 # define EVEN_DOUBLE	DOUBLE_X
238 #endif
239 	x0 = x->fp_mant[0];
240 	x1 = x->fp_mant[1];
241 	x2 = x->fp_mant[2];
242 	x3 = x->fp_mant[3];
243 	e = x->fp_exp;
244 	if (e & 1)		/* exponent is odd; use sqrt(2mant) */
245 		DOUBLE_X;
246 	/* THE FOLLOWING ASSUMES THAT RIGHT SHIFT DOES SIGN EXTENSION */
247 	x->fp_exp = e >> 1;	/* calculates (e&1 ? (e-1)/2 : e/2 */
248 
249 	/*
250 	 * Now calculate the mantissa root.  Since x is now in [1..4),
251 	 * we know that the first trip around the loop will definitely
252 	 * set the top bit in q, so we can do that manually and start
253 	 * the loop at the next bit down instead.  We must be sure to
254 	 * double x correctly while doing the `known q=1.0'.
255 	 *
256 	 * We do this one mantissa-word at a time, as noted above, to
257 	 * save work.  To avoid `(1U << 31) << 1', we also do the top bit
258 	 * outside of each per-word loop.
259 	 *
260 	 * The calculation `t = y + bit' breaks down into `t0 = y0, ...,
261 	 * t3 = y3, t? |= bit' for the appropriate word.  Since the bit
262 	 * is always a `new' one, this means that three of the `t?'s are
263 	 * just the corresponding `y?'; we use `#define's here for this.
264 	 * The variable `tt' holds the actual `t?' variable.
265 	 */
266 
267 	/* calculate q0 */
268 #define	t0 tt
269 	bit = FP_1;
270 	EVEN_DOUBLE;
271 	/* if (x >= (t0 = y0 | bit)) { */	/* always true */
272 		q = bit;
273 		x0 -= bit;
274 		y0 = bit << 1;
275 	/* } */
276 	ODD_DOUBLE;
277 	while ((bit >>= 1) != 0) {	/* for remaining bits in q0 */
278 		EVEN_DOUBLE;
279 		t0 = y0 | bit;		/* t = y + bit */
280 		if (x0 >= t0) {		/* if x >= t then */
281 			x0 -= t0;	/*	x -= t */
282 			q |= bit;	/*	q += bit */
283 			y0 |= bit << 1;	/*	y += bit << 1 */
284 		}
285 		ODD_DOUBLE;
286 	}
287 	x->fp_mant[0] = q;
288 #undef t0
289 
290 	/* calculate q1.  note (y0&1)==0. */
291 #define t0 y0
292 #define t1 tt
293 	q = 0;
294 	y1 = 0;
295 	bit = 1U << 31;
296 	EVEN_DOUBLE;
297 	t1 = bit;
298 	FPU_SUBS(d1, x1, t1);
299 	FPU_SUBC(d0, x0, t0);		/* d = x - t */
300 	if ((int)d0 >= 0) {		/* if d >= 0 (i.e., x >= t) then */
301 		x0 = d0, x1 = d1;	/*	x -= t */
302 		q = bit;		/*	q += bit */
303 		y0 |= 1;		/*	y += bit << 1 */
304 	}
305 	ODD_DOUBLE;
306 	while ((bit >>= 1) != 0) {	/* for remaining bits in q1 */
307 		EVEN_DOUBLE;		/* as before */
308 		t1 = y1 | bit;
309 		FPU_SUBS(d1, x1, t1);
310 		FPU_SUBC(d0, x0, t0);
311 		if ((int)d0 >= 0) {
312 			x0 = d0, x1 = d1;
313 			q |= bit;
314 			y1 |= bit << 1;
315 		}
316 		ODD_DOUBLE;
317 	}
318 	x->fp_mant[1] = q;
319 #undef t1
320 
321 	/* calculate q2.  note (y1&1)==0; y0 (aka t0) is fixed. */
322 #define t1 y1
323 #define t2 tt
324 	q = 0;
325 	y2 = 0;
326 	bit = 1U << 31;
327 	EVEN_DOUBLE;
328 	t2 = bit;
329 	FPU_SUBS(d2, x2, t2);
330 	FPU_SUBCS(d1, x1, t1);
331 	FPU_SUBC(d0, x0, t0);
332 	if ((int)d0 >= 0) {
333 		x0 = d0, x1 = d1, x2 = d2;
334 		q |= bit;
335 		y1 |= 1;		/* now t1, y1 are set in concrete */
336 	}
337 	ODD_DOUBLE;
338 	while ((bit >>= 1) != 0) {
339 		EVEN_DOUBLE;
340 		t2 = y2 | bit;
341 		FPU_SUBS(d2, x2, t2);
342 		FPU_SUBCS(d1, x1, t1);
343 		FPU_SUBC(d0, x0, t0);
344 		if ((int)d0 >= 0) {
345 			x0 = d0, x1 = d1, x2 = d2;
346 			q |= bit;
347 			y2 |= bit << 1;
348 		}
349 		ODD_DOUBLE;
350 	}
351 	x->fp_mant[2] = q;
352 #undef t2
353 
354 	/* calculate q3.  y0, t0, y1, t1 all fixed; y2, t2, almost done. */
355 #define t2 y2
356 #define t3 tt
357 	q = 0;
358 	y3 = 0;
359 	bit = 1U << 31;
360 	EVEN_DOUBLE;
361 	t3 = bit;
362 	FPU_SUBS(d3, x3, t3);
363 	FPU_SUBCS(d2, x2, t2);
364 	FPU_SUBCS(d1, x1, t1);
365 	FPU_SUBC(d0, x0, t0);
366 	if ((int)d0 >= 0) {
367 		x0 = d0, x1 = d1, x2 = d2, x3 = d3;
368 		q |= bit;
369 		y2 |= 1;
370 	}
371 	ODD_DOUBLE;
372 	while ((bit >>= 1) != 0) {
373 		EVEN_DOUBLE;
374 		t3 = y3 | bit;
375 		FPU_SUBS(d3, x3, t3);
376 		FPU_SUBCS(d2, x2, t2);
377 		FPU_SUBCS(d1, x1, t1);
378 		FPU_SUBC(d0, x0, t0);
379 		if ((int)d0 >= 0) {
380 			x0 = d0, x1 = d1, x2 = d2, x3 = d3;
381 			q |= bit;
382 			y3 |= bit << 1;
383 		}
384 		ODD_DOUBLE;
385 	}
386 	x->fp_mant[3] = q;
387 
388 	/*
389 	 * The result, which includes guard and round bits, is exact iff
390 	 * x is now zero; any nonzero bits in x represent sticky bits.
391 	 */
392 	x->fp_sticky = x0 | x1 | x2 | x3;
393 	return (x);
394 }
395