hodu_cpu_kernels 0.2.4

hodu cpu kernels
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

use hodu_cpu_kernels::*;

fn approx(v: Vec<f32>, digits: i32) -> Vec<f32> {
    let b = 10f32.powi(digits);
    v.iter().map(|t| f32::round(t * b) / b).collect()
}

#[test]
fn test_matmul_f32_2d() {
    // Matrix A: 2x3 = [[1, 2, 3], [4, 5, 6]]
    let lhs = [1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0];
    // Matrix B: 3x2 = [[7, 8], [9, 10], [11, 12]]
    let rhs = [7.0f32, 8.0, 9.0, 10.0, 11.0, 12.0];

    let m = 2;
    let k = 3;
    let n = 2;

    let lhs_ndim = 2;
    let rhs_ndim = 2;
    let batch_ndim = 0;

    let mut output = vec![0.0f32; m * n];
    let num_els = m * n;

    // Metadata for matmul:
    // [num_els, lhs_ndim, rhs_ndim, batch_ndim, lhs_shape, rhs_shape, batch_shape (empty),
    //  lhs_strides, rhs_strides, lhs_offset, rhs_offset, M, K, N]
    let lhs_shape = vec![m, k];
    let rhs_shape = vec![k, n];
    let batch_shape: Vec<usize> = vec![];
    let lhs_strides = vec![3, 1]; // row-major
    let rhs_strides = vec![2, 1]; // row-major
    let lhs_offset = 0;
    let rhs_offset = 0;

    let mut metadata = vec![num_els, lhs_ndim, rhs_ndim, batch_ndim];
    metadata.extend(&lhs_shape);
    metadata.extend(&rhs_shape);
    metadata.extend(&batch_shape);
    metadata.extend(&lhs_strides);
    metadata.extend(&rhs_strides);
    metadata.push(lhs_offset);
    metadata.push(rhs_offset);
    metadata.push(m);
    metadata.push(k);
    metadata.push(n);

    call_ops_matmul(
        matmul::F32,
        lhs.as_ptr() as *const core::ffi::c_void,
        rhs.as_ptr() as *const core::ffi::c_void,
        output.as_mut_ptr() as *mut core::ffi::c_void,
        &metadata,
    )
    .unwrap();

    // Expected: [[1*7+2*9+3*11, 1*8+2*10+3*12], [4*7+5*9+6*11, 4*8+5*10+6*12]]
    //         = [[58, 64], [139, 154]]
    assert_eq!(approx(output, 4), vec![58.0, 64.0, 139.0, 154.0]);
}

#[test]
fn test_matmul_f32_batch() {
    // Batch of 2 matrices:
    // Batch 0: A = [[1, 2], [3, 4]], B = [[5, 6], [7, 8]]
    // Batch 1: A = [[2, 3], [4, 5]], B = [[6, 7], [8, 9]]
    let lhs = [1.0f32, 2.0, 3.0, 4.0, 2.0, 3.0, 4.0, 5.0];
    let rhs = [5.0f32, 6.0, 7.0, 8.0, 6.0, 7.0, 8.0, 9.0];

    let batch_size = 2;
    let m = 2;
    let k = 2;
    let n = 2;

    let lhs_ndim = 3; // [batch, m, k]
    let rhs_ndim = 3; // [batch, k, n]
    let batch_ndim = 1;

    let mut output = vec![0.0f32; batch_size * m * n];
    let num_els = batch_size * m * n;

    // Metadata for batched matmul:
    let lhs_shape = vec![batch_size, m, k];
    let rhs_shape = vec![batch_size, k, n];
    let batch_shape = vec![batch_size];
    let lhs_strides = vec![4, 2, 1]; // batch stride = 4, row stride = 2, col stride = 1
    let rhs_strides = vec![4, 2, 1];
    let lhs_offset = 0;
    let rhs_offset = 0;

    let mut metadata = vec![num_els, lhs_ndim, rhs_ndim, batch_ndim];
    metadata.extend(&lhs_shape);
    metadata.extend(&rhs_shape);
    metadata.extend(&batch_shape);
    metadata.extend(&lhs_strides);
    metadata.extend(&rhs_strides);
    metadata.push(lhs_offset);
    metadata.push(rhs_offset);
    metadata.push(m);
    metadata.push(k);
    metadata.push(n);

    call_ops_matmul(
        matmul::F32,
        lhs.as_ptr() as *const core::ffi::c_void,
        rhs.as_ptr() as *const core::ffi::c_void,
        output.as_mut_ptr() as *mut core::ffi::c_void,
        &metadata,
    )
    .unwrap();

    // Expected:
    // Batch 0: [[1*5+2*7, 1*6+2*8], [3*5+4*7, 3*6+4*8]] = [[19, 22], [43, 50]]
    // Batch 1: [[2*6+3*8, 2*7+3*9], [4*6+5*8, 4*7+5*9]] = [[36, 41], [64, 73]]
    assert_eq!(approx(output, 4), vec![19.0, 22.0, 43.0, 50.0, 36.0, 41.0, 64.0, 73.0]);
}

#[test]
fn test_dot_f32_simple() {
    // Matrix A: 2x3 = [[1, 2, 3], [4, 5, 6]]
    let lhs = [1.0f32, 2.0, 3.0, 4.0, 5.0, 6.0];
    // Matrix B: 3x2 = [[7, 8], [9, 10], [11, 12]]
    let rhs = [7.0f32, 8.0, 9.0, 10.0, 11.0, 12.0];

    let m = 2; // rows of A
    let k = 3; // cols of A / rows of B
    let n = 2; // cols of B

    let mut output = vec![0.0f32; m * n];

    // Metadata: [M, K, N, lhs_stride_m, lhs_stride_k, rhs_stride_k, rhs_stride_n, lhs_offset, rhs_offset]
    let metadata = vec![m, k, n, 3, 1, 2, 1, 0, 0];

    call_ops_dot(
        dot::F32,
        lhs.as_ptr() as *const core::ffi::c_void,
        rhs.as_ptr() as *const core::ffi::c_void,
        output.as_mut_ptr() as *mut core::ffi::c_void,
        &metadata,
    )
    .unwrap();

    // Expected: [[1*7+2*9+3*11, 1*8+2*10+3*12], [4*7+5*9+6*11, 4*8+5*10+6*12]]
    //         = [[58, 64], [139, 154]]
    assert_eq!(approx(output, 4), vec![58.0, 64.0, 139.0, 154.0]);
}