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
|
// -*-C++-*-
#ifndef MATHFUNCS_FABS_H
#define MATHFUNCS_FABS_H
#include "mathfuncs_base.h"
#include <cmath>
namespace vecmathlib {
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_copysign(realvec_t x, realvec_t y) {
intvec_t value = as_int(x) & IV(U(~FP::signbit_mask));
intvec_t sign = as_int(y) & IV(FP::signbit_mask);
return as_float(sign | value);
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_fabs(realvec_t x) {
return as_float(as_int(x) & IV(U(~FP::signbit_mask)));
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_fdim(realvec_t x, realvec_t y) {
// return ifthen(x > y, x - y, RV(0.0));
return fmax(x - y, RV(0.0));
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_fma(realvec_t x, realvec_t y, realvec_t z) {
return x * y + z;
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_fmax(realvec_t x, realvec_t y) {
return ifthen(x < y, y, x);
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_fmin(realvec_t x, realvec_t y) {
return ifthen(y < x, y, x);
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_frexp(realvec_t x,
typename realvec_t::intvec_t *irp) {
intvec_t e = lsr(as_int(x) & IV(FP::exponent_mask), FP::mantissa_bits);
intvec_t ir = e - IV(FP::exponent_offset - 1);
ir = ifthen(convert_bool(e), ir, IV(std::numeric_limits<int_t>::min()));
#if defined VML_HAVE_INF
ir = ifthen(isinf(x), IV(std::numeric_limits<int_t>::max()), ir);
#endif
#if defined VML_HAVE_NAN
ir = ifthen(isnan(x), IV(std::numeric_limits<int_t>::min()), ir);
#endif
realvec_t r =
as_float((as_int(x) & IV(FP::signbit_mask | FP::mantissa_mask)) |
IV(FP::as_int(R(0.5)) & FP::exponent_mask));
boolvec_t iszero = x == RV(0.0);
ir = ifthen(iszero, IV(I(0)), ir);
r = ifthen(iszero, copysign(RV(R(0.0)), r), r);
*irp = ir;
return r;
}
template <typename realvec_t>
typename realvec_t::intvec_t mathfuncs<realvec_t>::vml_ilogb(realvec_t x) {
// TODO: Check SLEEF 2.80 algorithm
intvec_t e = lsr(as_int(x) & IV(FP::exponent_mask), FP::mantissa_bits);
intvec_t r = e - IV(FP::exponent_offset);
r = ifthen(convert_bool(e), r, IV(std::numeric_limits<int_t>::min()));
#if defined VML_HAVE_INF
r = ifthen(isinf(x), IV(std::numeric_limits<int_t>::max()), r);
#endif
#if defined VML_HAVE_NAN
r = ifthen(isnan(x), IV(std::numeric_limits<int_t>::min()), r);
#endif
return r;
}
template <typename realvec_t>
typename realvec_t::boolvec_t
mathfuncs<realvec_t>::vml_ieee_isfinite(realvec_t x) {
return (as_int(x) & IV(FP::exponent_mask)) != IV(FP::exponent_mask);
}
template <typename realvec_t>
typename realvec_t::boolvec_t
mathfuncs<realvec_t>::vml_ieee_isinf(realvec_t x) {
return (as_int(x) & IV(I(~FP::signbit_mask))) == IV(FP::exponent_mask);
}
template <typename realvec_t>
typename realvec_t::boolvec_t
mathfuncs<realvec_t>::vml_ieee_isnan(realvec_t x) {
return (as_int(x) & IV(FP::exponent_mask)) == IV(FP::exponent_mask) &&
(as_int(x) & IV(FP::mantissa_mask)) != IV(I(0));
}
template <typename realvec_t>
typename realvec_t::boolvec_t
mathfuncs<realvec_t>::vml_ieee_isnormal(realvec_t x) {
return (as_int(x) & IV(FP::exponent_mask)) != IV(FP::exponent_mask) &&
(as_int(x) & IV(FP::exponent_mask)) != IV(I(0));
}
template <typename realvec_t>
typename realvec_t::boolvec_t mathfuncs<realvec_t>::vml_isfinite(realvec_t x) {
#if defined VML_HAVE_INF || defined VML_HAVE_NAN
return vml_ieee_isfinite(x);
#else
return BV(true);
#endif
}
template <typename realvec_t>
typename realvec_t::boolvec_t mathfuncs<realvec_t>::vml_isinf(realvec_t x) {
#if defined VML_HAVE_INF
return vml_ieee_isinf(x);
#else
return BV(false);
#endif
}
template <typename realvec_t>
typename realvec_t::boolvec_t mathfuncs<realvec_t>::vml_isnan(realvec_t x) {
#if defined VML_HAVE_NAN
return vml_ieee_isnan(x);
#else
return BV(false);
#endif
}
template <typename realvec_t>
typename realvec_t::boolvec_t mathfuncs<realvec_t>::vml_isnormal(realvec_t x) {
#if defined VML_HAVE_DENORMALS || defined VML_HAVE_INF || defined VML_HAVE_NAN
return vml_ieee_isnormal(x);
#else
return BV(true);
#endif
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_ldexp(realvec_t x, intvec_t n) {
// TODO: Check SLEEF 2.80 algorithm
#if 0
realvec_t r = as_float(as_int(x) + (n << I(FP::mantissa_bits)));
r = ifthen((as_int(x) & IV(FP::exponent_mask)) == IV(I(0)), x, r);
return r;
#endif
realvec_t r = as_float(as_int(x) + (n << U(FP::mantissa_bits)));
int max_n = FP::max_exponent - FP::min_exponent;
boolvec_t underflow = n < IV(I(-max_n));
boolvec_t overflow = n > IV(I(max_n));
intvec_t old_exp = lsr(as_int(x) & IV(FP::exponent_mask), FP::mantissa_bits);
intvec_t new_exp = old_exp + n;
// TODO: check bit patterns instead
underflow =
underflow || new_exp < IV(I(FP::min_exponent + FP::exponent_offset));
overflow =
overflow || new_exp > IV(I(FP::max_exponent + FP::exponent_offset));
r = ifthen(underflow, copysign(RV(R(0.0)), x), r);
r = ifthen(overflow, copysign(RV(FP::infinity()), x), r);
boolvec_t dont_change = x == RV(R(0.0)) || isinf(x) || isnan(x);
r = ifthen(dont_change, x, r);
return r;
}
template <typename realvec_t>
realvec_t mathfuncs<realvec_t>::vml_mad(realvec_t x, realvec_t y, realvec_t z) {
return x * y + z;
}
template <typename realvec_t>
typename realvec_t::boolvec_t mathfuncs<realvec_t>::vml_signbit(realvec_t x) {
return convert_bool(as_int(x) & IV(FP::signbit_mask));
}
}; // namespace vecmathlib
#endif // #ifndef MATHFUNCS_FABS_H
|
