1 * $NetBSD: ssin.sa,v 1.4 2000/03/13 23:52:32 soren Exp $
3 * MOTOROLA MICROPROCESSOR & MEMORY TECHNOLOGY GROUP
4 * M68000 Hi-Performance Microprocessor Division
5 * M68040 Software Package
7 * M68040 Software Package Copyright (c) 1993, 1994 Motorola Inc.
10 * THE SOFTWARE is provided on an "AS IS" basis and without warranty.
11 * To the maximum extent permitted by applicable law,
12 * MOTOROLA DISCLAIMS ALL WARRANTIES WHETHER EXPRESS OR IMPLIED,
13 * INCLUDING IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A
14 * PARTICULAR PURPOSE and any warranty against infringement with
15 * regard to the SOFTWARE (INCLUDING ANY MODIFIED VERSIONS THEREOF)
16 * and any accompanying written materials.
18 * To the maximum extent permitted by applicable law,
19 * IN NO EVENT SHALL MOTOROLA BE LIABLE FOR ANY DAMAGES WHATSOEVER
20 * (INCLUDING WITHOUT LIMITATION, DAMAGES FOR LOSS OF BUSINESS
21 * PROFITS, BUSINESS INTERRUPTION, LOSS OF BUSINESS INFORMATION, OR
22 * OTHER PECUNIARY LOSS) ARISING OF THE USE OR INABILITY TO USE THE
23 * SOFTWARE. Motorola assumes no responsibility for the maintenance
24 * and support of the SOFTWARE.
26 * You are hereby granted a copyright license to use, modify, and
27 * distribute the SOFTWARE so long as this entire notice is retained
28 * without alteration in any modified and/or redistributed versions,
29 * and that such modified versions are clearly identified as such.
30 * No licenses are granted by implication, estoppel or otherwise
31 * under any patents or trademarks of Motorola, Inc.
36 * The entry point sSIN computes the sine of an input argument
37 * sCOS computes the cosine, and sSINCOS computes both. The
38 * corresponding entry points with a "d" computes the same
39 * corresponding function values for denormalized inputs.
41 * Input: Double-extended number X in location pointed to
42 * by address register a0.
44 * Output: The funtion value sin(X) or cos(X) returned in Fp0 if SIN or
45 * COS is requested. Otherwise, for SINCOS, sin(X) is returned
46 * in Fp0, and cos(X) is returned in Fp1.
48 * Modifies: Fp0 for SIN or COS; both Fp0 and Fp1 for SINCOS.
50 * Accuracy and Monotonicity: The returned result is within 1 ulp in
51 * 64 significant bit, i.e. within 0.5001 ulp to 53 bits if the
52 * result is subsequently rounded to double precision. The
53 * result is provably monotonic in double precision.
55 * Speed: The programs sSIN and sCOS take approximately 150 cycles for
56 * input argument X such that |X| < 15Pi, which is the usual
57 * situation. The speed for sSINCOS is approximately 190 cycles.
62 * 1. If SIN is invoked, set AdjN := 0; otherwise, set AdjN := 1.
64 * 2. If |X| >= 15Pi or |X| < 2**(-40), go to 7.
66 * 3. Decompose X as X = N(Pi/2) + r where |r| <= Pi/4. Let
67 * k = N mod 4, so in particular, k = 0,1,2,or 3. Overwirte
70 * 4. If k is even, go to 6.
72 * 5. (k is odd) Set j := (k-1)/2, sgn := (-1)**j. Return sgn*cos(r)
73 * where cos(r) is approximated by an even polynomial in r,
74 * 1 + r*r*(B1+s*(B2+ ... + s*B8)), s = r*r.
77 * 6. (k is even) Set j := k/2, sgn := (-1)**j. Return sgn*sin(r)
78 * where sin(r) is approximated by an odd polynomial in r
79 * r + r*s*(A1+s*(A2+ ... + s*A7)), s = r*r.
82 * 7. If |X| > 1, go to 9.
84 * 8. (|X|<2**(-40)) If SIN is invoked, return X; otherwise return 1.
86 * 9. Overwrite X by X := X rem 2Pi. Now that |X| <= Pi, go back to 3.
89 * 1. If |X| >= 15Pi or |X| < 2**(-40), go to 6.
91 * 2. Decompose X as X = N(Pi/2) + r where |r| <= Pi/4. Let
92 * k = N mod 4, so in particular, k = 0,1,2,or 3.
94 * 3. If k is even, go to 5.
96 * 4. (k is odd) Set j1 := (k-1)/2, j2 := j1 (EOR) (k mod 2), i.e.
97 * j1 exclusive or with the l.s.b. of k.
98 * sgn1 := (-1)**j1, sgn2 := (-1)**j2.
99 * SIN(X) = sgn1 * cos(r) and COS(X) = sgn2*sin(r) where
100 * sin(r) and cos(r) are computed as odd and even polynomials
101 * in r, respectively. Exit
103 * 5. (k is even) Set j1 := k/2, sgn1 := (-1)**j1.
104 * SIN(X) = sgn1 * sin(r) and COS(X) = sgn1*cos(r) where
105 * sin(r) and cos(r) are computed as odd and even polynomials
106 * in r, respectively. Exit
108 * 6. If |X| > 1, go to 8.
110 * 7. (|X|<2**(-40)) SIN(X) = X and COS(X) = 1. Exit.
112 * 8. Overwrite X by X := X rem 2Pi. Now that |X| <= Pi, go back to 2.
115 SSIN IDNT 2,1 Motorola 040 Floating Point Software Package
121 BOUNDS1 DC.L $3FD78000,$4004BC7E
122 TWOBYPI DC.L $3FE45F30,$6DC9C883
124 SINA7 DC.L $BD6AAA77,$CCC994F5
125 SINA6 DC.L $3DE61209,$7AAE8DA1
127 SINA5 DC.L $BE5AE645,$2A118AE4
128 SINA4 DC.L $3EC71DE3,$A5341531
130 SINA3 DC.L $BF2A01A0,$1A018B59,$00000000,$00000000
132 SINA2 DC.L $3FF80000,$88888888,$888859AF,$00000000
134 SINA1 DC.L $BFFC0000,$AAAAAAAA,$AAAAAA99,$00000000
136 COSB8 DC.L $3D2AC4D0,$D6011EE3
137 COSB7 DC.L $BDA9396F,$9F45AC19
139 COSB6 DC.L $3E21EED9,$0612C972
140 COSB5 DC.L $BE927E4F,$B79D9FCF
142 COSB4 DC.L $3EFA01A0,$1A01D423,$00000000,$00000000
144 COSB3 DC.L $BFF50000,$B60B60B6,$0B61D438,$00000000
146 COSB2 DC.L $3FFA0000,$AAAAAAAA,$AAAAAB5E
149 INVTWOPI DC.L $3FFC0000,$A2F9836E,$4E44152A
151 TWOPI1 DC.L $40010000,$C90FDAA2,$00000000,$00000000
152 TWOPI2 DC.L $3FDF0000,$85A308D4,$00000000,$00000000
179 *--SIN(X) = X FOR DENORMALIZED X
184 *--COS(X) = 1 FOR DENORMALIZED X
186 FMOVE.S #:3F800000,FP0
188 * 9D25B Fix: Sometimes the previous fmove.s sets fpsr bits
206 *--SAVE FPCR, FP1. CHECK IF |X| IS TOO SMALL OR LARGE
208 FMOVE.X (a0),FP0 ...LOAD INPUT
213 ANDI.L #$7FFFFFFF,D0 ...COMPACTIFY X
215 CMPI.L #$3FD78000,D0 ...|X| >= 2**(-40)?
220 CMPI.L #$4004BC7E,D0 ...|X| < 15 PI?
225 *--THIS IS THE USUAL CASE, |X| <= 15 PI.
226 *--THE ARGUMENT REDUCTION IS DONE BY TABLE LOOK UP.
228 FMUL.D TWOBYPI,FP1 ...X*2/PI
230 *--HIDE THE NEXT THREE INSTRUCTIONS
231 LEA PITBL+$200,A1 ...TABLE OF N*PI/2, N = -32,...,32
235 FMOVE.L FP1,N(a6) ...CONVERT TO INTEGER
239 ADDA.L D0,A1 ...A1 IS THE ADDRESS OF N*PIBY2
240 * ...WHICH IS IN TWO PIECES Y1 & Y2
242 FSUB.X (A1)+,FP0 ...X-Y1
244 FSUB.S (A1),FP0 ...FP0 IS R = (X-Y1)-Y2
247 *--continuation from REDUCEX
249 *--GET N+ADJN AND SEE IF SIN(R) OR COS(R) IS NEEDED
251 ADD.L ADJN(a6),D0 ...SEE IF D0 IS ODD OR EVEN
252 ROR.L #1,D0 ...D0 WAS ODD IFF D0 IS NEGATIVE
257 *--LET J BE THE LEAST SIG. BIT OF D0, LET SGN := (-1)**J.
258 *--THEN WE RETURN SGN*SIN(R). SGN*SIN(R) IS COMPUTED BY
259 *--R' + R'*S*(A1 + S(A2 + S(A3 + S(A4 + ... + SA7)))), WHERE
260 *--R' = SGN*R, S=R*R. THIS CAN BE REWRITTEN AS
261 *--R' + R'*S*( [A1+T(A3+T(A5+TA7))] + [S(A2+T(A4+TA6))])
263 *--NOTE THAT A3 THROUGH A7 ARE STORED IN DOUBLE PRECISION
264 *--WHILE A1 AND A2 ARE IN DOUBLE-EXTENDED FORMAT.
265 FMOVE.X FP0,X(a6) ...X IS R
266 FMUL.X FP0,FP0 ...FP0 IS S
267 *---HIDE THE NEXT TWO WHILE WAITING FOR FP0
272 FMUL.X FP1,FP1 ...FP1 IS T
273 *--HIDE THE NEXT TWO WHILE WAITING FOR FP1
277 * ...LEAST SIG. BIT OF D0 IN SIGN POSITION
278 EOR.L D0,X(a6) ...X IS NOW R'= SGN*R
280 FMUL.X FP1,FP3 ...TA7
281 FMUL.X FP1,FP2 ...TA6
283 FADD.D SINA5,FP3 ...A5+TA7
284 FADD.D SINA4,FP2 ...A4+TA6
286 FMUL.X FP1,FP3 ...T(A5+TA7)
287 FMUL.X FP1,FP2 ...T(A4+TA6)
289 FADD.D SINA3,FP3 ...A3+T(A5+TA7)
290 FADD.X SINA2,FP2 ...A2+T(A4+TA6)
292 FMUL.X FP3,FP1 ...T(A3+T(A5+TA7))
294 FMUL.X FP0,FP2 ...S(A2+T(A4+TA6))
295 FADD.X SINA1,FP1 ...A1+T(A3+T(A5+TA7))
296 FMUL.X X(a6),FP0 ...R'*S
298 FADD.X FP2,FP1 ...[A1+T(A3+T(A5+TA7))]+[S(A2+T(A4+TA6))]
299 *--FP3 RELEASED, RESTORE NOW AND TAKE SOME ADVANTAGE OF HIDING
300 *--FP2 RELEASED, RESTORE NOW AND TAKE FULL ADVANTAGE OF HIDING
303 FMUL.X FP1,FP0 ...SIN(R')-R'
306 FMOVE.L d1,FPCR ;restore users exceptions
307 FADD.X X(a6),FP0 ;last inst - possible exception set
312 *--LET J BE THE LEAST SIG. BIT OF D0, LET SGN := (-1)**J.
313 *--THEN WE RETURN SGN*COS(R). SGN*COS(R) IS COMPUTED BY
314 *--SGN + S'*(B1 + S(B2 + S(B3 + S(B4 + ... + SB8)))), WHERE
315 *--S=R*R AND S'=SGN*S. THIS CAN BE REWRITTEN AS
316 *--SGN + S'*([B1+T(B3+T(B5+TB7))] + [S(B2+T(B4+T(B6+TB8)))])
318 *--NOTE THAT B4 THROUGH B8 ARE STORED IN DOUBLE PRECISION
319 *--WHILE B2 AND B3 ARE IN DOUBLE-EXTENDED FORMAT, B1 IS -1/2
320 *--AND IS THEREFORE STORED AS SINGLE PRECISION.
322 FMUL.X FP0,FP0 ...FP0 IS S
323 *---HIDE THE NEXT TWO WHILE WAITING FOR FP0
328 FMUL.X FP1,FP1 ...FP1 IS T
329 *--HIDE THE NEXT TWO WHILE WAITING FOR FP1
330 FMOVE.X FP0,X(a6) ...X IS S
333 * ...LEAST SIG. BIT OF D0 IN SIGN POSITION
335 FMUL.X FP1,FP2 ...TB8
336 *--HIDE THE NEXT TWO WHILE WAITING FOR THE XU
337 EOR.L D0,X(a6) ...X IS NOW S'= SGN*S
340 FMUL.X FP1,FP3 ...TB7
341 *--HIDE THE NEXT TWO WHILE WAITING FOR THE XU
342 ORI.L #$3F800000,D0 ...D0 IS SGN IN SINGLE
343 MOVE.L D0,POSNEG1(a6)
345 FADD.D COSB6,FP2 ...B6+TB8
346 FADD.D COSB5,FP3 ...B5+TB7
348 FMUL.X FP1,FP2 ...T(B6+TB8)
349 FMUL.X FP1,FP3 ...T(B5+TB7)
351 FADD.D COSB4,FP2 ...B4+T(B6+TB8)
352 FADD.X COSB3,FP3 ...B3+T(B5+TB7)
354 FMUL.X FP1,FP2 ...T(B4+T(B6+TB8))
355 FMUL.X FP3,FP1 ...T(B3+T(B5+TB7))
357 FADD.X COSB2,FP2 ...B2+T(B4+T(B6+TB8))
358 FADD.S COSB1,FP1 ...B1+T(B3+T(B5+TB7))
360 FMUL.X FP2,FP0 ...S(B2+T(B4+T(B6+TB8)))
361 *--FP3 RELEASED, RESTORE NOW AND TAKE SOME ADVANTAGE OF HIDING
370 FMOVE.L d1,FPCR ;restore users exceptions
371 FADD.S POSNEG1(a6),FP0 ;last inst - possible exception set
376 *--IF |X| > 15PI, WE USE THE GENERAL ARGUMENT REDUCTION.
377 *--IF |X| < 2**(-40), RETURN X OR 1.
388 CLR.W XDCARE(a6) ...JUST IN CASE
389 FMOVE.L d1,FPCR ;restore users exceptions
390 FMOVE.X X(a6),FP0 ;last inst - possible exception set
395 FMOVE.S #:3F800000,FP0
397 FMOVE.L d1,FPCR ;restore users exceptions
398 FSUB.S #:00800000,FP0 ;last inst - possible exception set
403 *--WHEN REDUCEX IS USED, THE CODE WILL INEVITABLY BE SLOW.
404 *--THIS REDUCTION METHOD, HOWEVER, IS MUCH FASTER THAN USING
405 *--THE REMAINDER INSTRUCTION WHICH IS NOW IN SOFTWARE.
407 FMOVEM.X FP2-FP5,-(A7) ...save FP2 through FP5
409 FMOVE.S #:00000000,FP1
410 *--If compact form of abs(arg) in d0=$7ffeffff, argument is so large that
411 *--there is a danger of unwanted overflow in first LOOP iteration. In this
412 *--case, reduce argument by one remainder step to make subsequent reduction
414 cmpi.l #$7ffeffff,d0 ;is argument dangerously large?
416 move.l #$7ffe0000,FP_SCR2(a6) ;yes
417 * ;create 2**16383*PI/2
418 move.l #$c90fdaa2,FP_SCR2+4(a6)
420 ftst.x fp0 ;test sign of argument
421 move.l #$7fdc0000,FP_SCR3(a6) ;create low half of 2**16383*
423 move.l #$85a308d3,FP_SCR3+4(a6)
426 or.w #$8000,FP_SCR2(a6) ;positive arg
427 or.w #$8000,FP_SCR3(a6)
429 fadd.x FP_SCR2(a6),fp0 ;high part of reduction is exact
430 fmove.x fp0,fp1 ;save high result in fp1
431 fadd.x FP_SCR3(a6),fp0 ;low part of reduction
432 fsub.x fp0,fp1 ;determine low component of result
433 fadd.x FP_SCR3(a6),fp1 ;fp0/fp1 are reduced argument.
435 *--ON ENTRY, FP0 IS X, ON RETURN, FP0 IS X REM PI/2, |X| <= PI/4.
436 *--integer quotient will be stored in N
437 *--Intermeditate remainder is 66-bit long; (R,r) in (FP0,FP1)
440 FMOVE.X FP0,INARG(a6) ...+-2**K * F, 1 <= F < 2
442 MOVE.L D0,A1 ...save a copy of D0
444 SUBI.L #$00003FFF,D0 ...D0 IS K
448 SUBI.L #27,D0 ...D0 IS L := K-27
452 CLR.L D0 ...D0 IS L := 0
453 MOVE.L #1,ENDFLAG(a6)
456 *--FIND THE REMAINDER OF (R,r) W.R.T. 2**L * (PI/2). L IS SO CHOSEN
457 *--THAT INT( X * (2/PI) / 2**(L) ) < 2**29.
459 *--CREATE 2**(-L) * (2/PI), SIGN(INARG)*2**(63),
460 *--2**L * (PIby2_1), 2**L * (PIby2_2)
462 MOVE.L #$00003FFE,D2 ...BIASED EXPO OF 2/PI
463 SUB.L D0,D2 ...BIASED EXPO OF 2**(-L)*(2/PI)
465 MOVE.L #$A2F9836E,FP_SCR1+4(a6)
466 MOVE.L #$4E44152A,FP_SCR1+8(a6)
467 MOVE.W D2,FP_SCR1(a6) ...FP_SCR1 is 2**(-L)*(2/PI)
470 FMUL.X FP_SCR1(a6),FP2
471 *--WE MUST NOW FIND INT(FP2). SINCE WE NEED THIS VALUE IN
472 *--FLOATING POINT FORMAT, THE TWO FMOVE'S FMOVE.L FP <--> N
473 *--WILL BE TOO INEFFICIENT. THE WAY AROUND IT IS THAT
474 *--(SIGN(INARG)*2**63 + FP2) - SIGN(INARG)*2**63 WILL GIVE
475 *--US THE DESIRED VALUE IN FLOATING POINT.
477 *--HIDE SIX CYCLES OF INSTRUCTION
481 ORI.L #$5F000000,D2 ...D2 IS SIGN(INARG)*2**63 IN SGL
482 MOVE.L D2,TWOTO63(a6)
485 ADDI.L #$00003FFF,D2 ...BIASED EXPO OF 2**L * (PI/2)
488 FADD.S TWOTO63(a6),FP2 ...THE FRACTIONAL PART OF FP1 IS ROUNDED
490 *--HIDE 4 CYCLES OF INSTRUCTION; creating 2**(L)*Piby2_1 and 2**(L)*Piby2_2
491 MOVE.W D2,FP_SCR2(a6)
493 MOVE.L #$C90FDAA2,FP_SCR2+4(a6)
494 CLR.L FP_SCR2+8(a6) ...FP_SCR2 is 2**(L) * Piby2_1
497 FSUB.S TWOTO63(a6),FP2 ...FP2 is N
500 MOVE.W D0,FP_SCR3(a6)
502 MOVE.L #$85A308D3,FP_SCR3+4(a6)
503 CLR.L FP_SCR3+8(a6) ...FP_SCR3 is 2**(L) * Piby2_2
505 MOVE.L ENDFLAG(a6),D0
507 *--We are now ready to perform (R+r) - N*P1 - N*P2, P1 = 2**(L) * Piby2_1 and
508 *--P2 = 2**(L) * Piby2_2
510 FMul.X FP_SCR2(a6),FP4 ...W = N*P1
512 FMul.X FP_SCR3(a6),FP5 ...w = N*P2
514 *--we want P+p = W+w but |p| <= half ulp of P
515 *--Then, we need to compute A := R-P and a := r-p
516 FAdd.X FP5,FP3 ...FP3 is P
517 FSub.X FP3,FP4 ...W-P
519 FSub.X FP3,FP0 ...FP0 is A := R - P
520 FAdd.X FP5,FP4 ...FP4 is p = (W-P)+w
522 FMove.X FP0,FP3 ...FP3 A
523 FSub.X FP4,FP1 ...FP1 is a := r - p
525 *--Now we need to normalize (A,a) to "new (R,r)" where R+r = A+a but
526 *--|r| <= half ulp of R.
527 FAdd.X FP1,FP0 ...FP0 is R := A+a
528 *--No need to calculate r if this is the last loop
532 *--Need to calculate r
533 FSub.X FP0,FP3 ...A-R
534 FAdd.X FP3,FP1 ...FP1 is r := (A-R)+a
540 FMOVEM.X (A7)+,FP2-FP5
551 *--SIN AND COS OF X FOR DENORMALIZED X
553 FMOVE.S #:3F800000,FP1
554 bsr sto_cos ;store cosine result
562 FMOVE.X (a0),FP0 ...LOAD INPUT
567 ANDI.L #$7FFFFFFF,D0 ...COMPACTIFY X
569 CMPI.L #$3FD78000,D0 ...|X| >= 2**(-40)?
574 CMPI.L #$4004BC7E,D0 ...|X| < 15 PI?
580 *--THIS IS THE USUAL CASE, |X| <= 15 PI.
581 *--THE ARGUMENT REDUCTION IS DONE BY TABLE LOOK UP.
583 FMUL.D TWOBYPI,FP1 ...X*2/PI
585 *--HIDE THE NEXT THREE INSTRUCTIONS
586 LEA PITBL+$200,A1 ...TABLE OF N*PI/2, N = -32,...,32
590 FMOVE.L FP1,N(a6) ...CONVERT TO INTEGER
594 ADDA.L D0,A1 ...ADDRESS OF N*PIBY2, IN Y1, Y2
596 FSUB.X (A1)+,FP0 ...X-Y1
597 FSUB.S (A1),FP0 ...FP0 IS R = (X-Y1)-Y2
600 *--continuation point from REDUCEX
606 TST.L D0 ...D0 < 0 IFF N IS ODD
610 *--REGISTERS SAVED SO FAR: D0, A0, FP2.
612 FMOVE.X FP0,RPRIME(a6)
613 FMUL.X FP0,FP0 ...FP0 IS S = R*R
614 FMOVE.D SINA7,FP1 ...A7
615 FMOVE.D COSB8,FP2 ...B8
616 FMUL.X FP0,FP1 ...SA7
619 FMUL.X FP0,FP2 ...SB8
623 FADD.D SINA6,FP1 ...A6+SA7
626 FADD.D COSB7,FP2 ...B7+SB8
628 FMUL.X FP0,FP1 ...S(A6+SA7)
631 FMUL.X FP0,FP2 ...S(B7+SB8)
635 FADD.D SINA5,FP1 ...A5+S(A6+SA7)
636 MOVE.L #$3F800000,POSNEG1(a6)
638 FADD.D COSB6,FP2 ...B6+S(B7+SB8)
640 FMUL.X FP0,FP1 ...S(A5+S(A6+SA7))
641 FMUL.X FP0,FP2 ...S(B6+S(B7+SB8))
642 FMOVE.X FP0,SPRIME(a6)
644 FADD.D SINA4,FP1 ...A4+S(A5+S(A6+SA7))
646 FADD.D COSB5,FP2 ...B5+S(B6+S(B7+SB8))
648 FMUL.X FP0,FP1 ...S(A4+...)
649 FMUL.X FP0,FP2 ...S(B5+...)
651 FADD.D SINA3,FP1 ...A3+S(A4+...)
652 FADD.D COSB4,FP2 ...B4+S(B5+...)
654 FMUL.X FP0,FP1 ...S(A3+...)
655 FMUL.X FP0,FP2 ...S(B4+...)
657 FADD.X SINA2,FP1 ...A2+S(A3+...)
658 FADD.X COSB3,FP2 ...B3+S(B4+...)
660 FMUL.X FP0,FP1 ...S(A2+...)
661 FMUL.X FP0,FP2 ...S(B3+...)
663 FADD.X SINA1,FP1 ...A1+S(A2+...)
664 FADD.X COSB2,FP2 ...B2+S(B3+...)
666 FMUL.X FP0,FP1 ...S(A1+...)
667 FMUL.X FP2,FP0 ...S(B2+...)
671 FMUL.X RPRIME(a6),FP1 ...R'S(A1+...)
672 FADD.S COSB1,FP0 ...B1+S(B2...)
673 FMUL.X SPRIME(a6),FP0 ...S'(B1+S(B2+...))
675 move.l d1,-(sp) ;restore users mode & precision
676 andi.l #$ff,d1 ;mask off all exceptions
678 FADD.X RPRIME(a6),FP1 ...COS(X)
679 bsr sto_cos ;store cosine result
680 FMOVE.L (sp)+,FPCR ;restore users exceptions
681 FADD.S POSNEG1(a6),FP0 ...SIN(X)
687 *--REGISTERS SAVED SO FAR: FP2.
689 FMOVE.X FP0,RPRIME(a6)
690 FMUL.X FP0,FP0 ...FP0 IS S = R*R
691 FMOVE.D COSB8,FP1 ...B8
692 FMOVE.D SINA7,FP2 ...A7
693 FMUL.X FP0,FP1 ...SB8
694 FMOVE.X FP0,SPRIME(a6)
695 FMUL.X FP0,FP2 ...SA7
698 FADD.D COSB7,FP1 ...B7+SB8
699 FADD.D SINA6,FP2 ...A6+SA7
702 FMUL.X FP0,FP1 ...S(B7+SB8)
704 MOVE.L D0,POSNEG1(a6)
705 FMUL.X FP0,FP2 ...S(A6+SA7)
707 FADD.D COSB6,FP1 ...B6+S(B7+SB8)
708 FADD.D SINA5,FP2 ...A5+S(A6+SA7)
710 FMUL.X FP0,FP1 ...S(B6+S(B7+SB8))
711 FMUL.X FP0,FP2 ...S(A5+S(A6+SA7))
713 FADD.D COSB5,FP1 ...B5+S(B6+S(B7+SB8))
714 FADD.D SINA4,FP2 ...A4+S(A5+S(A6+SA7))
716 FMUL.X FP0,FP1 ...S(B5+...)
717 FMUL.X FP0,FP2 ...S(A4+...)
719 FADD.D COSB4,FP1 ...B4+S(B5+...)
720 FADD.D SINA3,FP2 ...A3+S(A4+...)
722 FMUL.X FP0,FP1 ...S(B4+...)
723 FMUL.X FP0,FP2 ...S(A3+...)
725 FADD.X COSB3,FP1 ...B3+S(B4+...)
726 FADD.X SINA2,FP2 ...A2+S(A3+...)
728 FMUL.X FP0,FP1 ...S(B3+...)
729 FMUL.X FP0,FP2 ...S(A2+...)
731 FADD.X COSB2,FP1 ...B2+S(B3+...)
732 FADD.X SINA1,FP2 ...A1+S(A2+...)
734 FMUL.X FP0,FP1 ...S(B2+...)
735 fmul.x fp2,fp0 ...s(a1+...)
739 FADD.S COSB1,FP1 ...B1+S(B2...)
740 FMUL.X RPRIME(a6),FP0 ...R'S(A1+...)
741 FMUL.X SPRIME(a6),FP1 ...S'(B1+S(B2+...))
743 move.l d1,-(sp) ;save users mode & precision
744 andi.l #$ff,d1 ;mask off all exceptions
746 FADD.S POSNEG1(a6),FP1 ...COS(X)
747 bsr sto_cos ;store cosine result
748 FMOVE.L (sp)+,FPCR ;restore users exceptions
749 FADD.X RPRIME(a6),FP0 ...SIN(X)
760 FMOVE.S #:3F800000,FP1
762 move.l d1,-(sp) ;save users mode & precision
763 andi.l #$ff,d1 ;mask off all exceptions
765 FSUB.S #:00800000,FP1
766 bsr sto_cos ;store cosine result
767 FMOVE.L (sp)+,FPCR ;restore users exceptions