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
|
//@ run-pass
#![allow(non_camel_case_types)]
//@ ignore-emscripten
// Test that the simd_reduce_{op} intrinsics produce the correct results.
#![feature(repr_simd, intrinsics)]
#[allow(non_camel_case_types)]
#[repr(simd)]
#[derive(Copy, Clone)]
struct i32x4(pub [i32; 4]);
#[repr(simd)]
#[derive(Copy, Clone)]
struct u32x4(pub [u32; 4]);
#[repr(simd)]
#[derive(Copy, Clone)]
struct f32x4(pub [f32; 4]);
#[repr(simd)]
#[derive(Copy, Clone)]
struct b8x4(pub [i8; 4]);
extern "rust-intrinsic" {
fn simd_reduce_add_unordered<T, U>(x: T) -> U;
fn simd_reduce_mul_unordered<T, U>(x: T) -> U;
fn simd_reduce_add_ordered<T, U>(x: T, acc: U) -> U;
fn simd_reduce_mul_ordered<T, U>(x: T, acc: U) -> U;
fn simd_reduce_min<T, U>(x: T) -> U;
fn simd_reduce_max<T, U>(x: T) -> U;
fn simd_reduce_and<T, U>(x: T) -> U;
fn simd_reduce_or<T, U>(x: T) -> U;
fn simd_reduce_xor<T, U>(x: T) -> U;
fn simd_reduce_all<T>(x: T) -> bool;
fn simd_reduce_any<T>(x: T) -> bool;
}
fn main() {
unsafe {
let x = i32x4([1, -2, 3, 4]);
let r: i32 = simd_reduce_add_unordered(x);
assert_eq!(r, 6_i32);
let r: i32 = simd_reduce_mul_unordered(x);
assert_eq!(r, -24_i32);
let r: i32 = simd_reduce_add_ordered(x, -1);
assert_eq!(r, 5_i32);
let r: i32 = simd_reduce_mul_ordered(x, -1);
assert_eq!(r, 24_i32);
let r: i32 = simd_reduce_min(x);
assert_eq!(r, -2_i32);
let r: i32 = simd_reduce_max(x);
assert_eq!(r, 4_i32);
let x = i32x4([-1, -1, -1, -1]);
let r: i32 = simd_reduce_and(x);
assert_eq!(r, -1_i32);
let r: i32 = simd_reduce_or(x);
assert_eq!(r, -1_i32);
let r: i32 = simd_reduce_xor(x);
assert_eq!(r, 0_i32);
let x = i32x4([-1, -1, 0, -1]);
let r: i32 = simd_reduce_and(x);
assert_eq!(r, 0_i32);
let r: i32 = simd_reduce_or(x);
assert_eq!(r, -1_i32);
let r: i32 = simd_reduce_xor(x);
assert_eq!(r, -1_i32);
}
unsafe {
let x = u32x4([1, 2, 3, 4]);
let r: u32 = simd_reduce_add_unordered(x);
assert_eq!(r, 10_u32);
let r: u32 = simd_reduce_mul_unordered(x);
assert_eq!(r, 24_u32);
let r: u32 = simd_reduce_add_ordered(x, 1);
assert_eq!(r, 11_u32);
let r: u32 = simd_reduce_mul_ordered(x, 2);
assert_eq!(r, 48_u32);
let r: u32 = simd_reduce_min(x);
assert_eq!(r, 1_u32);
let r: u32 = simd_reduce_max(x);
assert_eq!(r, 4_u32);
let t = u32::MAX;
let x = u32x4([t, t, t, t]);
let r: u32 = simd_reduce_and(x);
assert_eq!(r, t);
let r: u32 = simd_reduce_or(x);
assert_eq!(r, t);
let r: u32 = simd_reduce_xor(x);
assert_eq!(r, 0_u32);
let x = u32x4([t, t, 0, t]);
let r: u32 = simd_reduce_and(x);
assert_eq!(r, 0_u32);
let r: u32 = simd_reduce_or(x);
assert_eq!(r, t);
let r: u32 = simd_reduce_xor(x);
assert_eq!(r, t);
}
unsafe {
let x = f32x4([1., -2., 3., 4.]);
let r: f32 = simd_reduce_add_unordered(x);
assert_eq!(r, 6_f32);
let r: f32 = simd_reduce_mul_unordered(x);
assert_eq!(r, -24_f32);
let r: f32 = simd_reduce_add_ordered(x, 0.);
assert_eq!(r, 6_f32);
let r: f32 = simd_reduce_mul_ordered(x, 1.);
assert_eq!(r, -24_f32);
let r: f32 = simd_reduce_add_ordered(x, 1.);
assert_eq!(r, 7_f32);
let r: f32 = simd_reduce_mul_ordered(x, 2.);
assert_eq!(r, -48_f32);
let r: f32 = simd_reduce_min(x);
assert_eq!(r, -2_f32);
let r: f32 = simd_reduce_max(x);
assert_eq!(r, 4_f32);
}
unsafe {
let x = b8x4([!0, !0, !0, !0]);
let r: bool = simd_reduce_all(x);
assert_eq!(r, true);
let r: bool = simd_reduce_any(x);
assert_eq!(r, true);
let x = b8x4([!0, !0, 0, !0]);
let r: bool = simd_reduce_all(x);
assert_eq!(r, false);
let r: bool = simd_reduce_any(x);
assert_eq!(r, true);
let x = b8x4([0, 0, 0, 0]);
let r: bool = simd_reduce_all(x);
assert_eq!(r, false);
let r: bool = simd_reduce_any(x);
assert_eq!(r, false);
}
}
|
