brie-tree 0.1.2

A fast B+ Tree implementation that uses integer keys
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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159

use super::{SimdSearch, exact_div_unchecked};
use core::arch::aarch64::*;

impl SimdSearch for u8 {
    const SIMD_WIDTH: usize = 64;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_u8(search);
        let uint8x16x4_t(a, b, c, d) = unsafe { vld4q_u8(keys) };
        let a = vcgeq_u8(a, search);
        let b = vcgeq_u8(b, search);
        let c = vcgeq_u8(c, search);
        let d = vcgeq_u8(d, search);
        let ab = vsriq_n_u8(b, a, 1);
        let cd = vsriq_n_u8(d, c, 1);
        let abcd = vsriq_n_u8(cd, ab, 2);
        let abcd = vsriq_n_u8(abcd, abcd, 4);
        let abcd_low = vshrn_n_u16(vreinterpretq_u16_u8(abcd), 4);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        low.trailing_zeros() as usize
    }
}
impl SimdSearch for u16 {
    const SIMD_WIDTH: usize = 32;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_u16(search);
        let uint16x8x4_t(a, b, c, d) = unsafe { vld4q_u16(keys) };
        let a = vcgeq_u16(a, search);
        let b = vcgeq_u16(b, search);
        let c = vcgeq_u16(c, search);
        let d = vcgeq_u16(d, search);
        let ab = vsriq_n_u16(b, a, 2);
        let cd = vsriq_n_u16(d, c, 2);
        let abcd = vsriq_n_u16(cd, ab, 4);
        let abcd_low = vshrn_n_u16(abcd, 8);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        unsafe { exact_div_unchecked(low.trailing_zeros() as usize, 2) }
    }
}
impl SimdSearch for u32 {
    const SIMD_WIDTH: usize = 16;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_u32(search);
        let uint32x4x4_t(a, b, c, d) = unsafe { vld1q_u32_x4(keys) };
        let a = vcgeq_u32(a, search);
        let b = vcgeq_u32(b, search);
        let c = vcgeq_u32(c, search);
        let d = vcgeq_u32(d, search);
        let ab = vpmaxq_u16(vreinterpretq_u16_u32(a), vreinterpretq_u16_u32(b));
        let cd = vpmaxq_u16(vreinterpretq_u16_u32(c), vreinterpretq_u16_u32(d));
        let abcd = vpmaxq_u8(vreinterpretq_u8_u16(ab), vreinterpretq_u8_u16(cd));
        let abcd_low = vshrn_n_u16(vreinterpretq_u16_u8(abcd), 4);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        unsafe { exact_div_unchecked(low.trailing_zeros() as usize, 4) }
    }
}
impl SimdSearch for u64 {
    const SIMD_WIDTH: usize = 8;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_u64(search);
        let uint64x2x4_t(a, b, c, d) = unsafe { vld1q_u64_x4(keys) };
        let a = vcgeq_u64(a, search);
        let b = vcgeq_u64(b, search);
        let c = vcgeq_u64(c, search);
        let d = vcgeq_u64(d, search);
        let ab = vpmaxq_u32(vreinterpretq_u32_u64(a), vreinterpretq_u32_u64(b));
        let cd = vpmaxq_u32(vreinterpretq_u32_u64(c), vreinterpretq_u32_u64(d));
        let abcd = vpmaxq_u16(vreinterpretq_u16_u32(ab), vreinterpretq_u16_u32(cd));
        let abcd_low = vmovn_u16(abcd);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        unsafe { exact_div_unchecked(low.trailing_zeros() as usize, 8) }
    }
}
impl SimdSearch for u128 {}
impl SimdSearch for i8 {
    const SIMD_WIDTH: usize = 64;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_s8(search);
        let int8x16x4_t(a, b, c, d) = unsafe { vld4q_s8(keys) };
        let a = vcgeq_s8(a, search);
        let b = vcgeq_s8(b, search);
        let c = vcgeq_s8(c, search);
        let d = vcgeq_s8(d, search);
        let ab = vsriq_n_u8(b, a, 1);
        let cd = vsriq_n_u8(d, c, 1);
        let abcd = vsriq_n_u8(cd, ab, 2);
        let abcd = vsriq_n_u8(abcd, abcd, 4);
        let abcd_low = vshrn_n_u16(vreinterpretq_u16_u8(abcd), 4);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        low.trailing_zeros() as usize
    }
}
impl SimdSearch for i16 {
    const SIMD_WIDTH: usize = 32;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_s16(search);
        let int16x8x4_t(a, b, c, d) = unsafe { vld4q_s16(keys) };
        let a = vcgeq_s16(a, search);
        let b = vcgeq_s16(b, search);
        let c = vcgeq_s16(c, search);
        let d = vcgeq_s16(d, search);
        let ab = vsriq_n_u16(b, a, 2);
        let cd = vsriq_n_u16(d, c, 2);
        let abcd = vsriq_n_u16(cd, ab, 4);
        let abcd_low = vshrn_n_u16(abcd, 8);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        unsafe { exact_div_unchecked(low.trailing_zeros() as usize, 2) }
    }
}
impl SimdSearch for i32 {
    const SIMD_WIDTH: usize = 16;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_s32(search);
        let int32x4x4_t(a, b, c, d) = unsafe { vld1q_s32_x4(keys) };
        let a = vcgeq_s32(a, search);
        let b = vcgeq_s32(b, search);
        let c = vcgeq_s32(c, search);
        let d = vcgeq_s32(d, search);
        let ab = vpmaxq_u16(vreinterpretq_u16_u32(a), vreinterpretq_u16_u32(b));
        let cd = vpmaxq_u16(vreinterpretq_u16_u32(c), vreinterpretq_u16_u32(d));
        let abcd = vpmaxq_u8(vreinterpretq_u8_u16(ab), vreinterpretq_u8_u16(cd));
        let abcd_low = vshrn_n_u16(vreinterpretq_u16_u8(abcd), 4);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        unsafe { exact_div_unchecked(low.trailing_zeros() as usize, 4) }
    }
}
impl SimdSearch for i64 {
    const SIMD_WIDTH: usize = 8;
    #[inline]
    #[target_feature(enable = "neon")]
    unsafe fn simd_search(keys: *const Self, search: Self) -> usize {
        let search = vdupq_n_s64(search);
        let int64x2x4_t(a, b, c, d) = unsafe { vld1q_s64_x4(keys) };
        let a = vcgeq_s64(a, search);
        let b = vcgeq_s64(b, search);
        let c = vcgeq_s64(c, search);
        let d = vcgeq_s64(d, search);
        let ab = vpmaxq_u32(vreinterpretq_u32_u64(a), vreinterpretq_u32_u64(b));
        let cd = vpmaxq_u32(vreinterpretq_u32_u64(c), vreinterpretq_u32_u64(d));
        let abcd = vpmaxq_u16(vreinterpretq_u16_u32(ab), vreinterpretq_u16_u32(cd));
        let abcd_low = vmovn_u16(abcd);
        let low = vget_lane_u64(vreinterpret_u64_u8(abcd_low), 0);
        unsafe { exact_div_unchecked(low.trailing_zeros() as usize, 8) }
    }
}
impl SimdSearch for i128 {}