hannoy 0.1.3

HNSW Approximate Nearest Neighbors in Rust, based on LMDB and optimized for memory usage
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

use std::arch::aarch64::*;
use std::ptr::read_unaligned;

#[cfg(target_feature = "neon")]
use crate::unaligned_vector::UnalignedVector;

#[cfg(target_feature = "neon")]
pub(crate) unsafe fn euclid_similarity_neon(
    v1: &UnalignedVector<f32>,
    v2: &UnalignedVector<f32>,
) -> f32 {
    // We use the unaligned_float32x4_t helper function to read f32x4 NEON SIMD types
    // from potentially unaligned memory locations safely.
    // https://github.com/meilisearch/arroy/pull/13

    let n = v1.len();
    let m = n - (n % 16);
    let mut ptr1 = v1.as_ptr() as *const f32;
    let mut ptr2 = v2.as_ptr() as *const f32;
    let mut sum1 = vdupq_n_f32(0.);
    let mut sum2 = vdupq_n_f32(0.);
    let mut sum3 = vdupq_n_f32(0.);
    let mut sum4 = vdupq_n_f32(0.);

    let mut i: usize = 0;
    while i < m {
        let sub1 = vsubq_f32(unaligned_float32x4_t(ptr1), unaligned_float32x4_t(ptr2));
        sum1 = vfmaq_f32(sum1, sub1, sub1);

        let sub2 =
            vsubq_f32(unaligned_float32x4_t(ptr1.add(4)), unaligned_float32x4_t(ptr2.add(4)));
        sum2 = vfmaq_f32(sum2, sub2, sub2);

        let sub3 =
            vsubq_f32(unaligned_float32x4_t(ptr1.add(8)), unaligned_float32x4_t(ptr2.add(8)));
        sum3 = vfmaq_f32(sum3, sub3, sub3);

        let sub4 =
            vsubq_f32(unaligned_float32x4_t(ptr1.add(12)), unaligned_float32x4_t(ptr2.add(12)));
        sum4 = vfmaq_f32(sum4, sub4, sub4);

        ptr1 = ptr1.add(16);
        ptr2 = ptr2.add(16);
        i += 16;
    }
    let mut result = vaddvq_f32(sum1) + vaddvq_f32(sum2) + vaddvq_f32(sum3) + vaddvq_f32(sum4);
    for i in 0..n - m {
        let a = read_unaligned(ptr1.add(i));
        let b = read_unaligned(ptr2.add(i));
        result += (a - b).powi(2);
    }
    result
}

#[cfg(target_feature = "neon")]
pub(crate) unsafe fn dot_similarity_neon(
    v1: &UnalignedVector<f32>,
    v2: &UnalignedVector<f32>,
) -> f32 {
    // We use the unaligned_float32x4_t helper function to read f32x4 NEON SIMD types
    // from potentially unaligned memory locations safely.
    // https://github.com/meilisearch/arroy/pull/13

    let n = v1.len();
    let m = n - (n % 16);
    let mut ptr1 = v1.as_ptr() as *const f32;
    let mut ptr2 = v2.as_ptr() as *const f32;
    let mut sum1 = vdupq_n_f32(0.);
    let mut sum2 = vdupq_n_f32(0.);
    let mut sum3 = vdupq_n_f32(0.);
    let mut sum4 = vdupq_n_f32(0.);

    let mut i: usize = 0;
    while i < m {
        sum1 = vfmaq_f32(sum1, unaligned_float32x4_t(ptr1), unaligned_float32x4_t(ptr2));
        sum2 =
            vfmaq_f32(sum2, unaligned_float32x4_t(ptr1.add(4)), unaligned_float32x4_t(ptr2.add(4)));
        sum3 =
            vfmaq_f32(sum3, unaligned_float32x4_t(ptr1.add(8)), unaligned_float32x4_t(ptr2.add(8)));
        sum4 = vfmaq_f32(
            sum4,
            unaligned_float32x4_t(ptr1.add(12)),
            unaligned_float32x4_t(ptr2.add(12)),
        );
        ptr1 = ptr1.add(16);
        ptr2 = ptr2.add(16);
        i += 16;
    }
    let mut result = vaddvq_f32(sum1) + vaddvq_f32(sum2) + vaddvq_f32(sum3) + vaddvq_f32(sum4);
    for i in 0..n - m {
        let a = read_unaligned(ptr1.add(i));
        let b = read_unaligned(ptr2.add(i));
        result += a * b;
    }
    result
}

/// Reads 4xf32 in a stack-located array aligned on a f32 and reads a `float32x4_t` from it.
unsafe fn unaligned_float32x4_t(ptr: *const f32) -> float32x4_t {
    vld1q_f32(read_unaligned(ptr as *const [f32; 4]).as_ptr())
}

#[cfg(test)]
mod tests {
    #[cfg(target_feature = "neon")]
    #[test]
    fn test_spaces_neon() {
        use super::*;
        use crate::spaces::simple::*;

        if std::arch::is_aarch64_feature_detected!("neon") {
            let v1: Vec<f32> = vec![
                10., 11., 12., 13., 14., 15., 16., 17., 18., 19., 20., 21., 22., 23., 24., 25.,
                26., 27., 28., 29., 30., 31.,
            ];
            let v2: Vec<f32> = vec![
                40., 41., 42., 43., 44., 45., 46., 47., 48., 49., 50., 51., 52., 53., 54., 55.,
                56., 57., 58., 59., 60., 61.,
            ];

            let v1 = UnalignedVector::from_slice(&v1[..]);
            let v2 = UnalignedVector::from_slice(&v2[..]);

            let euclid_simd = unsafe { euclid_similarity_neon(&v1, &v2) };
            let euclid = euclidean_distance_non_optimized(&v1, &v2);
            assert_eq!(euclid_simd, euclid);

            let dot_simd = unsafe { dot_similarity_neon(&v1, &v2) };
            let dot = dot_product_non_optimized(&v1, &v2);
            assert_eq!(dot_simd, dot);

            // let cosine_simd = unsafe { cosine_preprocess_neon(v1.clone()) };
            // let cosine = cosine_preprocess(v1);
            // assert_eq!(cosine_simd, cosine);
        } else {
            println!("neon test skipped");
        }
    }
}