1 |
111 |
Agner |
/********************************* sincosf.as *******************************
|
2 |
|
|
* Author: Agner Fog
|
3 |
|
|
* date created: 2020-04-29
|
4 |
|
|
* Last modified: 2021-04-25
|
5 |
|
|
* Version: 1.11
|
6 |
|
|
* Project: ForwardCom library math.li
|
7 |
|
|
* Description: sin, cos, and tan functions. Calculate in radians, single precision
|
8 |
|
|
* The argument x can be a scalar or a vector
|
9 |
|
|
* The return value will be a vector with the same length
|
10 |
|
|
* C declaration: float sin(float x);
|
11 |
|
|
* C declaration: float cos(float x);
|
12 |
|
|
* C declaration: float tan(float x);
|
13 |
|
|
* C declaration: struct {float s; float c;} sincos(float x);
|
14 |
|
|
*
|
15 |
|
|
* This code is adapted from C++ vector class library www.github.com/vectorclass
|
16 |
|
|
* Copyright 2020-2021 GNU General Public License http://www.gnu.org/licenses
|
17 |
|
|
*****************************************************************************/
|
18 |
|
|
|
19 |
|
|
|
20 |
|
|
// define constants
|
21 |
|
|
% M_2_PI = 0.636619772367581343076 // 2./pi
|
22 |
|
|
|
23 |
|
|
% P0sinf = -1.6666654611E-1
|
24 |
|
|
% P1sinf = 8.3321608736E-3
|
25 |
|
|
% P2sinf = -1.9515295891E-4
|
26 |
|
|
% P0cosf = 4.166664568298827E-2
|
27 |
|
|
% P1cosf = -1.388731625493765E-3
|
28 |
|
|
% P2cosf = 2.443315711809948E-5
|
29 |
|
|
|
30 |
|
|
% DP1F = 0.78515625 * 2.
|
31 |
|
|
% DP2F = 2.4187564849853515625E-4 * 2.
|
32 |
|
|
% DP3F = 3.77489497744594108E-8 * 2.
|
33 |
|
|
|
34 |
|
|
|
35 |
|
|
code section execute align = 4
|
36 |
|
|
|
37 |
|
|
public _sinf: function, reguse = 0, 0x1BF
|
38 |
|
|
public _cosf: function, reguse = 0, 0x1BF
|
39 |
|
|
public _sincosf: function, reguse = 0, 0x1BF
|
40 |
|
|
public _tanf: function, reguse = 0, 0x1BF
|
41 |
|
|
|
42 |
|
|
// common entry for sin and sincos functions
|
43 |
|
|
_sinf function
|
44 |
|
|
_sincosf:
|
45 |
|
|
|
46 |
|
|
/* registers:
|
47 |
|
|
v0 = x
|
48 |
|
|
v1 = abs(x)
|
49 |
|
|
v1 = quadrant
|
50 |
|
|
v2 = x^2
|
51 |
|
|
v3 = x^3
|
52 |
|
|
v4 = x^4
|
53 |
|
|
v5 = temp
|
54 |
|
|
v5 = sin
|
55 |
|
|
v6 = unused (vacant flag for calling function)
|
56 |
|
|
v7 = cos
|
57 |
|
|
v8 = abs(x) reduced modulo pi/2
|
58 |
|
|
*/
|
59 |
|
|
|
60 |
|
|
// Find quadrant:
|
61 |
|
|
// 0 - pi/4 => 0
|
62 |
|
|
// pi/4 - 3*pi/4 => 1
|
63 |
|
|
// 3*pi/4 - 5*pi/4 => 2
|
64 |
|
|
// 5*pi/4 - 7*pi/4 => 3
|
65 |
|
|
// 7*pi/4 - 8*pi/4 => 4
|
66 |
|
|
|
67 |
|
|
// reduce modulo pi/2, with extended precision
|
68 |
|
|
//nop
|
69 |
|
|
float v1 = clear_bit(v0, 31) // abs(x)
|
70 |
|
|
float v5 = v1 * M_2_PI
|
71 |
|
|
float v5 = round(v5, 0) // round to integer
|
72 |
|
|
|
73 |
|
|
// x = ((xa - y * DP1) - y * DP2) - y * DP3;
|
74 |
|
|
float v8 = v5 * (-DP1F) + v1
|
75 |
|
|
float v8 = v5 * (-DP2F) + v8
|
76 |
|
|
float v8 = v5 * (-DP3F) + v8
|
77 |
|
|
|
78 |
|
|
float v3 = !(v5 > ((1 << 22) + 0.0)) // check for loss of precision and overflow, but not NAN
|
79 |
|
|
float v1 = v5 + ((1 << 23) + 0.0) // add magic number 2^23 to get integer into lowest bit
|
80 |
|
|
float v8 = v3 ? v8 : 0 // zero if out of range. result will be -1, 0, or 1
|
81 |
|
|
|
82 |
|
|
// Taylor expansion of sin and cos, valid for -pi/4 <= x <= pi/4
|
83 |
|
|
// s = polynomial_2(x^2, P0sinf, P1sinf, P2sinf) * (x*x^2) + x;
|
84 |
|
|
|
85 |
|
|
float v2 = v8 * v8 // x^2
|
86 |
|
|
float v3 = v8 * v2 // x^3
|
87 |
|
|
float v4 = v2 * v2 // x^4
|
88 |
|
|
float v5 = replace(v8, P0sinf) // broadcast to same length as x
|
89 |
|
|
float v5 = v2 * P1sinf + v5
|
90 |
|
|
float v5 = v4 * P2sinf + v5
|
91 |
|
|
float v5 = v5 * v3 + v8 // sin
|
92 |
|
|
|
93 |
|
|
// c = polynomial_2(x2, P0cosf, P1cosf, P2cosf) * (x2*x2) + nmul_add(0.5f, x2, 1.0f);
|
94 |
|
|
float v7 = replace(v8, P0cosf) // broadcast to same length as x
|
95 |
|
|
float v7 = v2 * P1cosf + v7
|
96 |
|
|
float v7 = v4 * P2cosf + v7
|
97 |
|
|
float v3 = replace(v8, 1.0)
|
98 |
|
|
float v3 = v2 * (-0.5) + v3 // 1 - 0.5*x^2
|
99 |
|
|
float v7 = v7 * v4 + v3 // cos
|
100 |
|
|
|
101 |
|
|
// swap sin and cos if odd quadrant
|
102 |
|
|
float v3 = v1 ? v7 : v5 // sin
|
103 |
|
|
float v4 = v1 ? v5 : v7 // cos
|
104 |
|
|
|
105 |
|
|
// get sign of sin
|
106 |
|
|
int32 v5 = v1 << 30 // get bit 1 into sign bit, x modulo pi/2 = 2 or 3
|
107 |
|
|
int32 v5 ^= v0 // toggle with sign of original x
|
108 |
|
|
int32 v5 = and(v5, 1 << 31) // isolate sign bit
|
109 |
|
|
float v0 = v3 ^ v5 // apply sign bit to sin
|
110 |
|
|
|
111 |
|
|
// get sign of cos
|
112 |
|
|
int32 v1 = v1 + 1 // change sign when x modulo pi/2 = 1 or 2
|
113 |
|
|
int32 v1 = v1 << 30 // get bit 1 into sign bit
|
114 |
|
|
int32 v1 = and(v1, 1 << 31) // isolate sign bit
|
115 |
|
|
float v1 = v4 ^ v1 // apply sign bit to cos
|
116 |
|
|
|
117 |
|
|
// return sin in v0, cos in v1
|
118 |
|
|
return
|
119 |
|
|
_sinf end
|
120 |
|
|
|
121 |
|
|
// cosine function
|
122 |
|
|
_cosf function
|
123 |
|
|
call _sincosf
|
124 |
|
|
float v0 = v1 // cos is in v1
|
125 |
|
|
return
|
126 |
|
|
_cosf end
|
127 |
|
|
|
128 |
|
|
// tangent function
|
129 |
|
|
_tanf function
|
130 |
|
|
call _sincosf
|
131 |
|
|
float v0 = v0 / v1 // tan(x) = sin(x)/cos(x)
|
132 |
|
|
return
|
133 |
|
|
_tanf end
|
134 |
|
|
|
135 |
|
|
code end
|