1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
|
/*
libffm - Free, pretty fast replacement for some math (libm) routines
on Linux/AXP, optimized for the 21164
Copyright (C) 1998 Joachim Wesner <joachim.wesner@frankfurt.netsurf.de>
and Kazushige Goto <goto@statabo.rim.or.jp>
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Library General Public
License as published by the Free Software Foundation; either
version 2 of the License, or (at your option) any later version.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Library General Public License for more details.
You should have received a copy of the GNU Library General Public
License along with this library (see file COPYING.LIB); if not, write
to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge,
MA 02139, USA.
*/
/*
Fast atan approximation including range reduction by Joachim Wesner,
<joachim.wesner@frankfurt.netsurf.de>, see also mc 8/1991 p. 78-93.
Done in June/July 1998.
*/
/*
improved and modified by Kazushige Goto <goto@statabo.rim.or.jp>
*/
/*
Now including full illegal argument handling by Joachim Wesner,
<joachim.wesner@frankfurt.netsurf.de>, November 1998
*/
.set noreorder
.set noat
#ifdef __ELF__
.section .rodata
#else
.rdata
#endif
.align 5
.type K,@object
K:
.t_floating 2.67949192431122695801e-1
.t_floating 1.73205080756887741522e0
.t_floating 5.77350269189625842081e-1
.t_floating 1.00000000000000000000e0
.t_floating 3.73205080756887719318e0
.t_floating 1.57079632679489655800e0
.t_floating 5.23598775598298815659e-1
.t_floating 1.04719755119659763132e0
.t_floating 4.43895157186999997356e-2
.t_floating 7.67936869066000032946e-2
.t_floating 1.11110978980510477498e-1
.t_floating 1.99999999987294480031e-1
.t_floating -6.48319351030299945160e-2
.t_floating -9.09037114191073997160e-2
.t_floating -1.42857141028255446580e-1
.t_floating -3.33333333333299286494e-1
.text
.align 5
.globl atan
.ent atan
atan:
lda $30, -16($30)
ldgp $29,.-atan($27)
stt $f16, 0($30)
#ifdef PROF
lda $28, _mcount
jsr $28, ($28), _mcount
unop
unop
#endif
.prologue 1
fabs $f16, $f17
lda $1,K
lda $2,0x7ff
sll $2, 52, $2
ldq $0, 0($30)
ldt $f18, 0($1) # K[0]
ldt $f19, 8($1) # K[1]
ldt $f20, 16($1) # K[2]
and $0,$2,$5
beq $5,$Denormal
cmpeq $2,$5,$4 ###
bne $4,$NaN_and_Inf ###
cmptlt $f18, $f17, $f18
ldt $f21, 24($1) # K[3]
mult $f19, $f17, $f14
ldt $f22, 32($1) # K[4]
ldt $f23, 64($1) # K[8]
cmptle $f17, $f21, $f13 # $f1 = (xa<=K[3])
lda $3, 40($1) # c_addr = &K[5]
mult $f20, $f17, $f12
cmptle $f17, $f22, $f22
ldt $f24, 72($1) # K[9]
fclr $f0 # corr = 0.0
ldt $f25, 80($1) # K[10]
ldt $f26, 88($1) # K[11]
ldt $f27, 96($1) # K[12]
ldt $f28,104($1) # K[13]
fbeq $f18, $L6
fneg $f17, $f15 # y7 = -xa
fbeq $f13, $L7
addt $f19,$f17,$f15 # y7 = K[1] + xa
lda $3, 48($1) # c_addr = &K[6]
subt $f14,$f21,$f21 # y5 = y3 - y5
br $31,$L8
$L7:
fbeq $f22,$L8
addt $f20,$f17,$f15 # y7 = K[2] + xa
lda $3, 56($1) # c_addr = &K[7]
subt $f12,$f21,$f21 # y5 = y4 - y5
$L8:
divt $f21,$f15,$f17 # xa = y5/y7
ldt $f0, 0($3) # load corr
.align 4
$L6:
ldt $f29,112($1) # K[14]
mult $f17, $f17, $f10 # y1 = xa * xa
ldt $f30,120($1) # K[15]
mult $f10, $f10, $f11 # y2 = y1 * y1
mult $f23, $f10, $f23 # r1 = K[8] * y1
mult $f24, $f10, $f24 # r2 = K[9] * y1
mult $f25, $f10, $f25 # r3 = K[10] * y1
mult $f10, $f11, $f12 # y3 = y1 * y2
mult $f11, $f11, $f13 # y4 = y2 * y2
addt $f23, $f27, $f23 # r1 = r1 + K[12]
addt $f24, $f28, $f24 # r2 = r2 + K[13]
mult $f26, $f10, $f26 # r4 = K[11] * y1
addt $f25, $f29, $f25 # r3 = r3 + K[14]
mult $f11, $f12, $f14 # y5 = y3 * y2
mult $f12, $f13, $f15 # y7 = y4 * y3
addt $f26, $f30, $f26 # r4 = r4 + K[15]
mult $f25, $f12, $f25 # r3 = r3 * y3
mult $f24, $f14, $f24 # r2 = r2 * y5
mult $f23, $f15, $f23 # r1 = r1 * y7
mult $f26, $f10, $f26 # r4 = r4 * y1
addt $f23, $f24, $f23 # r1 = r1 + r2
addt $f25, $f26, $f25 # r3 = r3 + r4
addt $f23, $f25, $f23 # r1 = r1 + r3
mult $f23, $f17, $f23 # r1 = r1 * xa
addt $f23, $f17, $f23 # r1 = r1 + xa
addt $f23, $f0, $f0 # r1 = r1 + corr
addq $30,16,$30 ###
fblt $f16, $L12
ret
$L12:
fneg $f0, $f0
ret
.align 4
$Denormal:
fmov $f16,$f0
lda $30,16($30)
ret
.align 4
$NaN_and_Inf:
andnot $0,$2,$3
sll $3,1,$3
bne $3,$NaN
ldt $f0,40($1)
cpys $f16,$f0,$f0
lda $30,16($30)
ret
.align 4
$NaN:
lda $0,-1
stq $0, 0($30)
ldt $f0, 0($30)
lda $30,16($30)
ret
.end atan
|
