fearless_simd 0.4.0

Safer and easier SIMD
Documentation 
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

// Copyright 2024 the Fearless_SIMD Authors
// SPDX-License-Identifier: Apache-2.0 OR MIT

//! Access to SSE intrinsics.

use crate::impl_macros::delegate;
#[cfg(target_arch = "x86")]
use core::arch::x86 as arch;
#[cfg(target_arch = "x86_64")]
use core::arch::x86_64 as arch;

use arch::*;

/// A token for SSE intrinsics on `x86` and `x86_64`.
#[derive(Clone, Copy, Debug)]
pub struct Sse {
    _private: (),
}

#[expect(
    clippy::missing_safety_doc,
    reason = "TODO: https://github.com/linebender/fearless_simd/issues/40"
)]
impl Sse {
    /// Create a SIMD token.
    ///
    /// # Safety
    ///
    /// The required CPU features must be available.
    #[inline]
    pub const unsafe fn new_unchecked() -> Self {
        Self { _private: () }
    }

    delegate! { arch:
        fn _mm_add_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_add_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_sub_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_sub_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_mul_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_mul_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_div_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_div_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_sqrt_ss(a: __m128) -> __m128;
        fn _mm_sqrt_ps(a: __m128) -> __m128;
        fn _mm_rcp_ss(a: __m128) -> __m128;
        fn _mm_rcp_ps(a: __m128) -> __m128;
        fn _mm_rsqrt_ss(a: __m128) -> __m128;
        fn _mm_rsqrt_ps(a: __m128) -> __m128;
        fn _mm_min_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_min_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_max_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_max_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_and_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_andnot_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_or_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_xor_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpeq_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmplt_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmple_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpgt_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpge_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpneq_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpnlt_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpnle_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpngt_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpnge_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpord_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpunord_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpeq_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmplt_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmple_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpgt_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpge_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpneq_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpnlt_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpnle_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpngt_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpnge_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpord_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_cmpunord_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_comieq_ss(a: __m128, b: __m128) -> i32;
        fn _mm_comilt_ss(a: __m128, b: __m128) -> i32;
        fn _mm_comile_ss(a: __m128, b: __m128) -> i32;
        fn _mm_comigt_ss(a: __m128, b: __m128) -> i32;
        fn _mm_comige_ss(a: __m128, b: __m128) -> i32;
        fn _mm_comineq_ss(a: __m128, b: __m128) -> i32;
        fn _mm_ucomieq_ss(a: __m128, b: __m128) -> i32;
        fn _mm_ucomilt_ss(a: __m128, b: __m128) -> i32;
        fn _mm_ucomile_ss(a: __m128, b: __m128) -> i32;
        fn _mm_ucomigt_ss(a: __m128, b: __m128) -> i32;
        fn _mm_ucomige_ss(a: __m128, b: __m128) -> i32;
        fn _mm_ucomineq_ss(a: __m128, b: __m128) -> i32;
        fn _mm_cvtss_si32(a: __m128) -> i32;
        fn _mm_cvt_ss2si(a: __m128) -> i32;
        fn _mm_cvttss_si32(a: __m128) -> i32;
        fn _mm_cvtt_ss2si(a: __m128) -> i32;
        fn _mm_cvtss_f32(a: __m128) -> f32;
        fn _mm_cvtsi32_ss(a: __m128, b: i32) -> __m128;
        fn _mm_cvt_si2ss(a: __m128, b: i32) -> __m128;
        fn _mm_set_ss(a: f32) -> __m128;
        fn _mm_set1_ps(a: f32) -> __m128;
        fn _mm_set_ps1(a: f32) -> __m128;
        fn _mm_set_ps(a: f32, b: f32, c: f32, d: f32) -> __m128;
        fn _mm_setr_ps(a: f32, b: f32, c: f32, d: f32) -> __m128;
        fn _mm_setzero_ps() -> __m128;
        fn _mm_shuffle_ps<const MASK: i32>(a: __m128, b: __m128) -> __m128;
        fn _mm_unpackhi_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_unpacklo_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_movehl_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_movelh_ps(a: __m128, b: __m128) -> __m128;
        fn _mm_movemask_ps(a: __m128) -> i32;
        unsafe fn _mm_load_ss(p: *const f32) -> __m128;
        unsafe fn _mm_load1_ps(p: *const f32) -> __m128;
        unsafe fn _mm_load_ps1(p: *const f32) -> __m128;
        unsafe fn _mm_load_ps(p: *const f32) -> __m128;
        unsafe fn _mm_loadu_ps(p: *const f32) -> __m128;
        unsafe fn _mm_loadr_ps(p: *const f32) -> __m128;
        unsafe fn _mm_loadu_si64(mem_addr: *const u8) -> __m128i;
        unsafe fn _mm_store_ss(p: *mut f32, a: __m128);
        unsafe fn _mm_store1_ps(p: *mut f32, a: __m128);
        unsafe fn _mm_store_ps1(p: *mut f32, a: __m128);
        unsafe fn _mm_store_ps(p: *mut f32, a: __m128);
        unsafe fn _mm_storeu_ps(p: *mut f32, a: __m128);
        unsafe fn _mm_storer_ps(p: *mut f32, a: __m128);
        fn _mm_move_ss(a: __m128, b: __m128) -> __m128;
        fn _mm_sfence();
        #[expect(clippy::not_unsafe_ptr_arg_deref, reason="Prefetch has no preconditions, so is valid to accept a pointer.")]
        fn _mm_prefetch<const STRATEGY: i32>(p: *const i8);
        fn _mm_undefined_ps() -> __m128;
        #[allow(non_snake_case)]
        fn _MM_TRANSPOSE4_PS(
            row0: &mut __m128,
            row1: &mut __m128,
            row2: &mut __m128,
            row3: &mut __m128,
        );
        unsafe fn _mm_stream_ps(mem_addr: *mut f32, a: __m128);
    }
}