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
|
/*
v4f.h
Copyright 2011-14 Tim Goetze <tim@quitte.de>
http://quitte.de/dsp/
Vectors of four 32-bit floats for SIMD arithmetic, gcc style.
Implementation is evolving so the code is not maximally elegant yet, sorry.
*/
/*
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 3
of the License, or (at your option) any later version.
This program 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 General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
02111-1307, USA or point your web browser to http://www.gnu.org.
*/
#ifndef DSP_V4F_H
#define DSP_V4F_H
#ifdef __SSE__
#include <xmmintrin.h>
#endif
/* caution, gcc apparently doesnt always honour the alignment: segfault */
typedef float v4f_t __attribute__ ((vector_size(16), aligned(16)));
inline v4f_t v4f (float x)
{ v4f_t v = {x,x,x,x}; return v; }
inline v4f_t v4f (float x0, float x1, float x2, float x3)
{ v4f_t v = {x0,x1,x2,x3}; return v; }
inline v4f_t v4f (float *x)
{ v4f_t v = {x[0],x[1],x[2],x[3]}; return v; }
#define v4fa(x) ((float *) &x)
/* gcc's __builtin_shuffle is useless */
#ifdef __SSE__
#define v4f_shuffle(x,s3,s2,s1,s0) \
_mm_shuffle_ps(x,x,((s0)<<6|(s1)<<4|(s2)<<2|s3))
#define v4f_shuffle2(x,y,s3,s2,s1,s0) \
_mm_shuffle_ps(x,y,((s0)<<6|(s1)<<4|(s2)<<2|s3))
#define v4f_max(x,y) _mm_max_ps(x,y)
#define v4f_min(x,y) _mm_min_ps(x,y)
#else
inline v4f_t v4f_shuffle(v4f_t x, int s3, int s2, int s1, int s0)
{
v4f_t y=x;
float*px=(float*)&x; float*py=(float*)&y;
px[3]=py[s0]; px[2]=py[s1]; px[1]=py[s2]; px[0]=py[s3];
return x;
}
#endif
inline float v4f_sum (v4f_t v)
{
float * f = (float *) &v;
return f[0]+f[1]+f[2]+f[3];
}
/* mapping a float to float function [e.g. sinf() e.a.] to a vector */
typedef float (*f2f_fn) (float f);
template <f2f_fn fn>
v4f_t v4f_map (v4f_t x)
{
v4f_t y;
float * s = (float *) &x;
float * d = (float *) &y;
for (uint i = 0; i < 4; ++i)
d[i] = fn(s[i]);
return y;
}
#define v4f_0 ((v4f_t) {0, 0, 0, 0})
#define v4f_half ((v4f_t) {.5, .5, .5, .5})
#define v4f_1 ((v4f_t) {1, 1, 1, 1})
#define v4f_2 ((v4f_t) {2, 2, 2, 2})
#define v4f_halfpi ((v4f_t) {.5*M_PI, .5*M_PI, .5*M_PI, .5*M_PI})
#define v4f_pi ((v4f_t) {M_PI, M_PI, M_PI, M_PI})
#define v4f_2pi ((v4f_t) {2*M_PI, 2*M_PI, 2*M_PI, 2*M_PI})
template <int N>
class V4fArray
{
private:
/* sufficient space to align actual array to 16-byte boundary */
char _data [(N+1) * sizeof(v4f_t)];
public:
v4f_t * v;
V4fArray()
{
uint64 p = (uint64) ((uint64) _data + 16) & ~15ll;
v = (v4f_t*) p;
}
void reset() { memset (_data, 0, sizeof (_data)); }
v4f_t & operator [] (int i) {return v[i];}
};
class V4fData
{
private:
void * _data;
public:
v4f_t * v;
V4fData() {_data = 0;}
~V4fData() {free(_data);}
operator v4f_t * () {return v;}
operator void * () {return (void *) v;}
void init (int N)
{
_data = calloc(N+1, sizeof(v4f_t));
uint64 p = (16 + (uint64) _data) & ~15ll;
v = (v4f_t*) p;
}
};
/* four parallel sine oscillators */
class Sin4f
{
public:
char _data [4 * sizeof (v4f_t)];
int z;
public:
/* get 16-byte aligned data base */
inline v4f_t * data()
{
uint64 p = ((uint64) _data + 16) & ~15ll;
return (v4f_t *) p;
}
inline void set_f (v4f_t f)
{
v4f_t *y = data();
v4f_t w = -v4f_pi * f;
y[0] = v4f_map<__builtin_sinf> (w);
y[1] = v4f_map<__builtin_sinf> (v4f_2 * w);
/* b in above scalar implementation is y[2] in the flat data */
y[2] = v4f_2 * v4f_map<__builtin_cosf> (w); /* b */
z = 0;
}
inline v4f_t get()
{
v4f_t *y = data();
register v4f_t s = y[2] * y[z];
z ^= 1;
s -= y[z];
return y[z] = s;
}
};
#endif /* DSP_V4F_H */
|
