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
|
/******************************************************
*
* zexy - implementation file
*
* (c) 2006 Tim Blechmann
*
*
******************************************************
*
* license: GNU General Public License v.2
*
******************************************************/
#include "zexySIMD.h"
typedef struct _absgn
{
t_object x_obj;
t_float x_f;
} t_absgn;
/* ------------------------ sigABSGN~ ----------------------------- */
static t_class *sigABSGN_class;
static t_int *sigABSGN_perform(t_int *w)
{
t_sample *in = (t_sample *)(w[1]);
t_sample *out = (t_sample *)(w[2]);
t_sample *out2 = (t_sample *)(w[3]);
int n = (int)(w[4]);
while (n--)
{
t_sample val = *in++;
*out++ = fabsf(val);
if (val>0.) *out2++=1.;
else if (val<0.) *out2++=-1.;
else *out2++=0.;
}
return (w+5);
}
#ifdef __SSE__
static int l_bitmask[] ={0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff};
static int l_sgnbitmask[]={0x80000000, 0x80000000, 0x80000000, 0x80000000};
static t_int *sigABSGN_performSSE(t_int *w)
{
__m128 *in = (__m128 *)(w[1]);
__m128 *out1 = (__m128 *)(w[2]);
__m128 *out2 = (__m128 *)(w[3]);
int n = (int)(w[4])>>3;
const __m128 bitmask= _mm_loadu_ps((float*)l_bitmask);
const __m128 sgnmask= _mm_loadu_ps((float*)l_sgnbitmask);
const __m128 zero = _mm_setzero_ps();
const __m128 one = _mm_set1_ps(1.f);
do {
__m128 val, val2, xmm0, xmm1, xmm2, xmm3;
val=in[0];
xmm0 = _mm_cmpneq_ps(val, zero); /* mask for non-zeros */
xmm1 = _mm_and_ps (val, sgnmask);/* sign (without value) */
xmm0 = _mm_and_ps (xmm0, one); /* (abs) value: (val==0.f)?0.f:1.f */
out1[0]= _mm_and_ps (val, bitmask);/* abs: set sign-bit to "+" */
out2[0]= _mm_or_ps (xmm1, xmm0); /* merge sign and value */
val2=in[1];
xmm2 = _mm_cmpneq_ps(val2, zero); /* mask for non-zeros */
xmm3 = _mm_and_ps (val2, sgnmask);/* sign (without value) */
xmm2 = _mm_and_ps (xmm2, one); /* (abs) value: (val==0.f)?0.f:1.f */
out1[1]= _mm_and_ps (val2, bitmask);/* abs: set sign-bit to "+" */
out2[1]= _mm_or_ps (xmm3, xmm2); /* merge sign and value */
in +=2;
out1+=2;
out2+=2;
}
while (--n);
return (w+5);
}
#endif /* __SSE__ */
static void sigABSGN_dsp(t_absgn *x, t_signal **sp)
{
ZEXY_USEVAR(x);
#ifdef __SSE__
if(
ZEXY_TYPE_EQUAL(t_sample, float) && /* currently SSE2 code is only for float (not for double) */
Z_SIMD_CHKBLOCKSIZE(sp[0]->s_n) &&
Z_SIMD_CHKALIGN(sp[0]->s_vec) &&
Z_SIMD_CHKALIGN(sp[1]->s_vec) &&
Z_SIMD_CHKALIGN(sp[2]->s_vec) &&
zexy_testSSE(sigABSGN_perform, sigABSGN_performSSE, 1, 2)
)
{
dsp_add(sigABSGN_performSSE, 4, sp[0]->s_vec, sp[1]->s_vec, sp[2]->s_vec, sp[0]->s_n);
} else
#endif
{
dsp_add(sigABSGN_perform, 4, sp[0]->s_vec, sp[1]->s_vec, sp[2]->s_vec, sp[0]->s_n);
}
}
static void sigABSGN_helper(void)
{
post("\n%c absgn~ \t\t:: absolute value and sign of a signal", HEARTSYMBOL);
post(" \t\t copyright (c) Tim Blechmann 2006");
}
static void *sigABSGN_new(void)
{
t_absgn *x = (t_absgn *)pd_new(sigABSGN_class);
x->x_f=0.f;
outlet_new(&x->x_obj, gensym("signal"));
outlet_new(&x->x_obj, gensym("signal"));
return (x);
}
void absgn_tilde_setup(void)
{
sigABSGN_class = class_new(gensym("absgn~"), (t_newmethod)sigABSGN_new, 0,
sizeof(t_absgn), 0, A_DEFFLOAT, 0);
CLASS_MAINSIGNALIN(sigABSGN_class, t_absgn, x_f);
class_addmethod(sigABSGN_class, (t_method)sigABSGN_dsp, gensym("dsp"), 0);
class_addmethod(sigABSGN_class, (t_method)sigABSGN_helper, gensym("help"), 0);
class_sethelpsymbol(sigABSGN_class, gensym("zigbinops"));
zexy_register("absgn~");
}
|
