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
|
/******************************************************
*
* zexy - implementation file
*
* copyleft (c) IOhannes m zmölnig
*
* 1999:forum::für::umläute:2005
*
* institute of electronic music and acoustics (iem)
*
******************************************************
*
* license: GNU General Public License v.2
*
******************************************************/
/*
abs~: absolute value of signal
2112:forum::für::umläute:2005
*/
#include "zexy.h"
typedef struct _abs
{
t_object x_obj;
float x_f;
} t_abs;
/* ------------------------ sigABS~ ----------------------------- */
static t_class *sigABS_class;
static t_int *sigABS_perform(t_int *w)
{
t_float *in = (t_float *)(w[1]);
t_float *out = (t_float *)(w[2]);
int n = (int)(w[3]);
while (n--) *out++ = fabsf(*in++);
return (w+4);
}
#ifdef __SSE__
static long l_bitmask[]={0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff};
static t_int *sigABS_performSSE(t_int *w)
{
__m128 *in = (__m128 *)(w[1]);
__m128 *out = (__m128 *)(w[2]);
int n = (int)(w[3])>>4;
__m128 bitmask= _mm_loadu_ps((float*)l_bitmask);
while (n--) {
out[0] = _mm_and_ps(in[0] , bitmask);
out[1] = _mm_and_ps(in[1] , bitmask);
out[2] = _mm_and_ps(in[2] , bitmask);
out[3] = _mm_and_ps(in[3] , bitmask);
in +=4;
out+=4;
}
#if 0
/*
* handwritten SSE-code by tim blechmann
*
* JMZ: the above (using intrinsics) is a little bit slower
* but still about 4* as fast as the generic code
* i prefer using intrinsics as i don't have to learn how to
* assemble
*/
asm(
".section .rodata \n"
".align 16 \n"
"2: \n"
".long 2147483647 \n" /* bitmask */
".long 2147483647 \n" /* 0x7fffffff */
".long 2147483647 \n"
".long 2147483647 \n"
".text \n"
"movaps (2b), %%xmm0 \n" /* xmm0 = bitmask */
"shrl $4, %2 \n"
/* loop: *dest = abs(*src) */
"1: \n"
"movaps (%0,%3), %%xmm1 \n"
"andps %%xmm0, %%xmm1 \n"
"movaps %%xmm1, (%1,%3) \n"
"movaps 16(%0,%3), %%xmm2 \n"
"andps %%xmm0, %%xmm2 \n"
"movaps %%xmm2, 16(%1,%3) \n"
"movaps 32(%0,%3), %%xmm3 \n"
"andps %%xmm0, %%xmm3 \n"
"movaps %%xmm3, 32(%1,%3) \n"
"movaps 48(%0,%3), %%xmm4 \n"
"andps %%xmm0, %%xmm4 \n"
"movaps %%xmm4, 48(%1,%3) \n"
"addl $64, %3 \n"
"loop 1b \n"
:
:"r"(in), "r"(out), "c"(n), "r"(0)
:"%xmm0", "%xmm1", "%xmm2", "%xmm3", "%xmm4"
);
#endif /*0*/
return (w+4);
}
#endif /* __SSE__ */
static void sigABS_dsp(t_abs *x, t_signal **sp)
{
#ifdef __SSE__
if(
Z_SIMD_CHKBLOCKSIZE(sp[0]->s_n)&&
Z_SIMD_CHKALIGN(sp[0]->s_vec)&&
Z_SIMD_CHKALIGN(sp[1]->s_vec))
{
dsp_add(sigABS_performSSE, 3, sp[0]->s_vec, sp[1]->s_vec, sp[0]->s_n);
} else
#endif
{
dsp_add(sigABS_perform, 3, sp[0]->s_vec, sp[1]->s_vec, sp[0]->s_n);
}
}
static void sigABS_helper(void)
{
post("\n%c abs~ \t\t:: absolute value of a signal", HEARTSYMBOL);
}
static void *sigABS_new(void)
{
t_abs *x = (t_abs *)pd_new(sigABS_class);
x->x_f=0.f;
outlet_new(&x->x_obj, gensym("signal"));
return (x);
}
void abs_tilde_setup(void)
{
sigABS_class = class_new(gensym("abs~"), (t_newmethod)sigABS_new, 0,
sizeof(t_abs), 0, A_DEFFLOAT, 0);
CLASS_MAINSIGNALIN(sigABS_class, t_abs, x_f);
class_addmethod(sigABS_class, (t_method)sigABS_dsp, gensym("dsp"), 0);
class_addmethod(sigABS_class, (t_method)sigABS_helper, gensym("help"), 0);
class_sethelpsymbol(sigABS_class, gensym("zexy/sigbinops+"));
zexy_register("abs~");
}
|
