numr 0.5.2

High-performance numerical computing with multi-backend GPU acceleration (CPU/CUDA/WebGPU)
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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247

//! SIMD-accelerated comparison operations
//!
//! This module provides AVX2 and AVX-512 implementations for element-wise
//! comparison operations (eq, ne, lt, le, gt, ge).
//!
//! # SIMD Approach
//!
//! - Use SIMD compare intrinsics to generate masks
//! - Blend between 1.0 and 0.0 vectors based on masks
//! - Output 1.0 for true, 0.0 for false (matching scalar behavior)

#[cfg(target_arch = "x86_64")]
mod avx2;
#[cfg(target_arch = "x86_64")]
mod avx512;

#[cfg(target_arch = "aarch64")]
mod aarch64;

use super::{SimdLevel, detect_simd};
use crate::ops::CompareOp;

/// Minimum length to justify SIMD overhead
const SIMD_THRESHOLD: usize = 32;

/// SIMD comparison for f32
///
/// # Safety
/// - `a`, `b`, and `out` must point to `len` elements
#[inline]
pub unsafe fn compare_f32(op: CompareOp, a: *const f32, b: *const f32, out: *mut f32, len: usize) {
    let level = detect_simd();

    if len < SIMD_THRESHOLD || level == SimdLevel::Scalar {
        compare_scalar_f32(op, a, b, out, len);
        return;
    }

    #[cfg(target_arch = "x86_64")]
    match level {
        SimdLevel::Avx512 => avx512::compare_f32(op, a, b, out, len),
        SimdLevel::Avx2Fma => avx2::compare_f32(op, a, b, out, len),
        _ => compare_scalar_f32(op, a, b, out, len),
    }

    #[cfg(target_arch = "aarch64")]
    match level {
        SimdLevel::Neon | SimdLevel::NeonFp16 => aarch64::neon::compare_f32(op, a, b, out, len),
        _ => compare_scalar_f32(op, a, b, out, len),
    }

    #[cfg(not(any(target_arch = "x86_64", target_arch = "aarch64")))]
    compare_scalar_f32(op, a, b, out, len);
}

/// SIMD comparison for f64
///
/// # Safety
/// - `a`, `b`, and `out` must point to `len` elements
#[inline]
pub unsafe fn compare_f64(op: CompareOp, a: *const f64, b: *const f64, out: *mut f64, len: usize) {
    let level = detect_simd();

    if len < SIMD_THRESHOLD || level == SimdLevel::Scalar {
        compare_scalar_f64(op, a, b, out, len);
        return;
    }

    #[cfg(target_arch = "x86_64")]
    match level {
        SimdLevel::Avx512 => avx512::compare_f64(op, a, b, out, len),
        SimdLevel::Avx2Fma => avx2::compare_f64(op, a, b, out, len),
        _ => compare_scalar_f64(op, a, b, out, len),
    }

    #[cfg(target_arch = "aarch64")]
    match level {
        SimdLevel::Neon | SimdLevel::NeonFp16 => aarch64::neon::compare_f64(op, a, b, out, len),
        _ => compare_scalar_f64(op, a, b, out, len),
    }

    #[cfg(not(any(target_arch = "x86_64", target_arch = "aarch64")))]
    compare_scalar_f64(op, a, b, out, len);
}

// ============================================================================
// Scalar fallbacks
// ============================================================================

/// Scalar comparison for f32
#[inline]
pub unsafe fn compare_scalar_f32(
    op: CompareOp,
    a: *const f32,
    b: *const f32,
    out: *mut f32,
    len: usize,
) {
    match op {
        CompareOp::Eq => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) == *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Ne => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) != *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Lt => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) < *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Le => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) <= *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Gt => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) > *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Ge => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) >= *b.add(i) { 1.0 } else { 0.0 };
            }
        }
    }
}

/// Scalar comparison for f64
#[inline]
pub unsafe fn compare_scalar_f64(
    op: CompareOp,
    a: *const f64,
    b: *const f64,
    out: *mut f64,
    len: usize,
) {
    match op {
        CompareOp::Eq => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) == *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Ne => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) != *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Lt => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) < *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Le => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) <= *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Gt => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) > *b.add(i) { 1.0 } else { 0.0 };
            }
        }
        CompareOp::Ge => {
            for i in 0..len {
                *out.add(i) = if *a.add(i) >= *b.add(i) { 1.0 } else { 0.0 };
            }
        }
    }
}

half_binary_op!(compare, compare_f32, CompareOp);

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_compare_eq_f32() {
        let a = [1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0];
        let b = [1.0f32, 3.0, 3.0, 5.0, 5.0, 6.0, 8.0, 8.0];
        let mut out = [0.0f32; 8];
        let mut out_ref = [0.0f32; 8];

        unsafe {
            compare_f32(CompareOp::Eq, a.as_ptr(), b.as_ptr(), out.as_mut_ptr(), 8);
            compare_scalar_f32(
                CompareOp::Eq,
                a.as_ptr(),
                b.as_ptr(),
                out_ref.as_mut_ptr(),
                8,
            );
        }

        assert_eq!(out, out_ref);
        assert_eq!(out, [1.0, 0.0, 1.0, 0.0, 1.0, 1.0, 0.0, 1.0]);
    }

    #[test]
    fn test_compare_lt_f32() {
        let a = [1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0];
        let b = [2.0f32, 2.0, 2.0, 5.0, 4.0, 7.0, 7.0, 9.0];
        let mut out = [0.0f32; 8];
        let mut out_ref = [0.0f32; 8];

        unsafe {
            compare_f32(CompareOp::Lt, a.as_ptr(), b.as_ptr(), out.as_mut_ptr(), 8);
            compare_scalar_f32(
                CompareOp::Lt,
                a.as_ptr(),
                b.as_ptr(),
                out_ref.as_mut_ptr(),
                8,
            );
        }

        assert_eq!(out, out_ref);
        assert_eq!(out, [1.0, 0.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0]);
    }

    #[test]
    fn test_compare_large_f32() {
        let len = 256;
        let a: Vec<f32> = (0..len).map(|x| x as f32).collect();
        let b: Vec<f32> = (0..len).map(|x| (x as f32) - 0.5).collect();
        let mut out = vec![0.0f32; len];
        let mut out_ref = vec![0.0f32; len];

        unsafe {
            compare_f32(CompareOp::Gt, a.as_ptr(), b.as_ptr(), out.as_mut_ptr(), len);
            compare_scalar_f32(
                CompareOp::Gt,
                a.as_ptr(),
                b.as_ptr(),
                out_ref.as_mut_ptr(),
                len,
            );
        }

        assert_eq!(out, out_ref);
    }
}