llml 0.2.3

Implementation of basic math data types with high level frontend and low level backend
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

arm_use!();
use crate::{vec::{EucVecf4, EucVecf2}, mat::Matd2, others::Zero};
use std::{ops::{Add, Sub, Mul, Div, Neg}};

use super::EucVecd4;

#[derive(Clone, Copy, PartialEq, Eq)]
#[repr(transparent)]
pub struct Matf2 (pub(crate) EucVecf4);
impl_mat2!(Matf2, f32);

impl Matf2 {
    #[inline(always)]
    pub fn new (a: [f32;4]) -> Self {
        Self(EucVecf4::new(a))
    }

    #[inline(always)]
    pub fn x (&self) -> EucVecf2 {
        unsafe { EucVecf2(vget_low_f32(self.0.0)) }
    }

    #[inline(always)]
    pub fn xx (&self) -> f32 {
        self.0.x()
    }

    #[inline(always)]
    pub fn xy (&self) -> f32 {
        self.0.y()
    }

    #[inline(always)]
    pub fn y (&self) -> EucVecf2 {
        unsafe { EucVecf2(vget_high_f32(self.0.0)) }
    }

    #[inline(always)]
    pub fn yx (&self) -> f32 {
        self.0.z()
    }

    #[inline(always)]
    pub fn yy (&self) -> f32 {
        self.0.w()
    }

    #[inline(always)]
    pub fn tr (self) -> f32 {
        self.0.x() + self.0.w()
    }

    #[inline(always)]
    pub fn det (self) -> f32 {
        unsafe {
            let v1 = vget_low_f32(self.0.0);
            let v2 = vrev64_f32(vget_high_f32(self.0.0));
            let m1 = EucVecf2(vmul_f32(v1, v2));

            m1.x() - m1.y()
        }
    }

    #[inline(always)]
    pub fn inv (self) -> Option<Self> {
        let det = self.det();
        if det.is_zero() {
            return None
        }

        let vec = -EucVecf2::new([self.0.y(), self.0.z()]);
        Some(Self(EucVecf4::new([self.0.w(), vec.x(), vec.y(), self.0.x()]) / det))
    }

    #[inline(always)]
    pub unsafe fn inv_unsafe (self) -> Self {
        let vec = -EucVecf2::new([self.0.y(), self.0.z()]);
        Self(EucVecf4::new([self.0.w(), vec.x(), vec.y(), self.0.x()]) / self.det())
    }
}

impl Mul<EucVecf2> for Matf2 {
    type Output = EucVecf2;

    #[inline(always)]
    fn mul(self, rhs: EucVecf2) -> Self::Output {
        unsafe {
            let mul = vmulq_f32(self.0.0, vcombine_f32(rhs.0, rhs.0));
            let v1 = vtrn2q_f32(mul, mul);
            let v2 = vtrn1q_f32(mul, mul);

            let res = vaddq_f32(v1, v2);
            EucVecf2(vget_low_f32(vuzp2q_f32(res, res)))
        }
    }
}

impl Mul for Matf2 {
    type Output = Matf2;

    #[inline(always)]
    fn mul (self, rhs: Self) -> Self::Output {
        unsafe {
            let v1 = vtrn1q_f32(self.0.0, self.0.0);
            let v2 = vget_low_f32(rhs.0.0);
            let v2 = vcombine_f32(v2, v2);
            let m1 = vmulq_f32(v1, v2);

            let v3 = vtrn2q_f32(self.0.0, self.0.0);
            let v4 = vget_high_f32(rhs.0.0);
            let v4 = vcombine_f32(v4, v4);
            let m2 = vmulq_f32(v3, v4);

            Self(EucVecf4(vaddq_f32(m1, m2)))
        }
    }
}

impl Neg for Matf2 {
    type Output = Matf2;

    #[inline(always)]
    fn neg(self) -> Self::Output {
        Self(-self.0)
    }
}

impl Into<[f32;4]> for Matf2 {
    #[inline(always)]
    fn into(self) -> [f32;4] {
        self.0.into()
    }
}

impl Into<Matd2> for Matf2 {
    #[inline(always)]
    fn into(self) -> Matd2 {
        Matd2(Into::<EucVecd4>::into(self.0))
    }
}