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
|
/*
* Vector math abstractions.
*
* Copyright (c) 2019-2023, Arm Limited.
* SPDX-License-Identifier: MIT OR Apache-2.0 WITH LLVM-exception
*/
#ifndef _V_MATH_H
#define _V_MATH_H
#if !__aarch64__
# error "Cannot build without AArch64"
#endif
#define VPCS_ATTR __attribute__ ((aarch64_vector_pcs))
#define V_NAME_F1(fun) _ZGVnN4v_##fun##f
#define V_NAME_D1(fun) _ZGVnN2v_##fun
#define V_NAME_F2(fun) _ZGVnN4vv_##fun##f
#define V_NAME_D2(fun) _ZGVnN2vv_##fun
#include <stdint.h>
#include "../math_config.h"
#include <arm_neon.h>
/* Shorthand helpers for declaring constants. */
# define V2(X) { X, X }
# define V4(X) { X, X, X, X }
# define V8(X) { X, X, X, X, X, X, X, X }
static inline int
v_any_u16h (uint16x4_t x)
{
return vget_lane_u64 (vreinterpret_u64_u16 (x), 0) != 0;
}
static inline int
v_lanes32 (void)
{
return 4;
}
static inline float32x4_t
v_f32 (float x)
{
return (float32x4_t) V4 (x);
}
static inline uint32x4_t
v_u32 (uint32_t x)
{
return (uint32x4_t) V4 (x);
}
/* true if any elements of a v_cond result is non-zero. */
static inline int
v_any_u32 (uint32x4_t x)
{
/* assume elements in x are either 0 or -1u. */
return vpaddd_u64 (vreinterpretq_u64_u32 (x)) != 0;
}
static inline int
v_any_u32h (uint32x2_t x)
{
return vget_lane_u64 (vreinterpret_u64_u32 (x), 0) != 0;
}
static inline float32x4_t
v_lookup_f32 (const float *tab, uint32x4_t idx)
{
return (float32x4_t){tab[idx[0]], tab[idx[1]], tab[idx[2]], tab[idx[3]]};
}
static inline uint32x4_t
v_lookup_u32 (const uint32_t *tab, uint32x4_t idx)
{
return (uint32x4_t){tab[idx[0]], tab[idx[1]], tab[idx[2]], tab[idx[3]]};
}
static inline float32x4_t
v_call_f32 (float (*f) (float), float32x4_t x, float32x4_t y, uint32x4_t p)
{
return (float32x4_t){p[0] ? f (x[0]) : y[0], p[1] ? f (x[1]) : y[1],
p[2] ? f (x[2]) : y[2], p[3] ? f (x[3]) : y[3]};
}
static inline float32x4_t
v_call2_f32 (float (*f) (float, float), float32x4_t x1, float32x4_t x2,
float32x4_t y, uint32x4_t p)
{
return (float32x4_t){p[0] ? f (x1[0], x2[0]) : y[0],
p[1] ? f (x1[1], x2[1]) : y[1],
p[2] ? f (x1[2], x2[2]) : y[2],
p[3] ? f (x1[3], x2[3]) : y[3]};
}
static inline int
v_lanes64 (void)
{
return 2;
}
static inline float64x2_t
v_f64 (double x)
{
return (float64x2_t) V2 (x);
}
static inline uint64x2_t
v_u64 (uint64_t x)
{
return (uint64x2_t) V2 (x);
}
/* true if any elements of a v_cond result is non-zero. */
static inline int
v_any_u64 (uint64x2_t x)
{
/* assume elements in x are either 0 or -1u. */
return vpaddd_u64 (x) != 0;
}
static inline float64x2_t
v_lookup_f64 (const double *tab, uint64x2_t idx)
{
return (float64x2_t){tab[idx[0]], tab[idx[1]]};
}
static inline uint64x2_t
v_lookup_u64 (const uint64_t *tab, uint64x2_t idx)
{
return (uint64x2_t){tab[idx[0]], tab[idx[1]]};
}
static inline float64x2_t
v_call_f64 (double (*f) (double), float64x2_t x, float64x2_t y, uint64x2_t p)
{
double p1 = p[1];
double x1 = x[1];
if (likely (p[0]))
y[0] = f (x[0]);
if (likely (p1))
y[1] = f (x1);
return y;
}
#endif
|
