numrs2 0.3.3

A Rust implementation inspired by NumPy for numerical computing (NumRS2)
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

// Reduction operations for f64 arrays

// Reduction parameters
struct ReductionParams {
    op_type: u32,
    array_size: u32,
    workgroup_size: u32,
    _padding: u32,
}

// Bindings
@group(0) @binding(0) var<storage, read> input: array<f64>;
@group(0) @binding(1) var<storage, read_write> output: array<f64>;
@group(0) @binding(2) var<uniform> params: ReductionParams;

// Shared memory for workgroup reduction
var<workgroup> shared_data: array<f64, 256>;

@compute @workgroup_size(256)
fn sum(@builtin(global_invocation_id) global_id: vec3<u32>, @builtin(local_invocation_id) local_id: vec3<u32>, @builtin(workgroup_id) workgroup_id: vec3<u32>) {
    let global_idx = global_id.x;
    let local_idx = local_id.x;

    // Load data into shared memory
    var value: f64 = 0.0;
    if (global_idx < params.array_size) {
        value = input[global_idx];
    }
    shared_data[local_idx] = value;

    workgroupBarrier();

    // Reduction in shared memory
    var stride = params.workgroup_size / 2u;
    while (stride > 0u) {
        if (local_idx < stride) {
            shared_data[local_idx] = shared_data[local_idx] + shared_data[local_idx + stride];
        }
        workgroupBarrier();
        stride = stride / 2u;
    }

    // Write the result to the output buffer
    if (local_idx == 0u) {
        output[workgroup_id.x] = shared_data[0];
    }
}

@compute @workgroup_size(256)
fn mean(@builtin(global_invocation_id) global_id: vec3<u32>, @builtin(local_invocation_id) local_id: vec3<u32>, @builtin(workgroup_id) workgroup_id: vec3<u32>) {
    // First compute the sum using the same algorithm as above
    let global_idx = global_id.x;
    let local_idx = local_id.x;

    // Load data into shared memory
    var value: f64 = 0.0;
    if (global_idx < params.array_size) {
        value = input[global_idx];
    }
    shared_data[local_idx] = value;

    workgroupBarrier();

    // Reduction in shared memory
    var stride = params.workgroup_size / 2u;
    while (stride > 0u) {
        if (local_idx < stride) {
            shared_data[local_idx] = shared_data[local_idx] + shared_data[local_idx + stride];
        }
        workgroupBarrier();
        stride = stride / 2u;
    }

    // Write the result to the output buffer
    if (local_idx == 0u) {
        // Calculate the mean
        let sum = shared_data[0];
        let count = f64(min(params.workgroup_size, params.array_size - workgroup_id.x * params.workgroup_size));
        output[workgroup_id.x] = sum / count;
    }
}

@compute @workgroup_size(256)
fn max(@builtin(global_invocation_id) global_id: vec3<u32>, @builtin(local_invocation_id) local_id: vec3<u32>, @builtin(workgroup_id) workgroup_id: vec3<u32>) {
    let global_idx = global_id.x;
    let local_idx = local_id.x;

    // Load data into shared memory
    var value: f64 = -1.7976931348623157e+308; // -DBL_MAX
    if (global_idx < params.array_size) {
        value = input[global_idx];
    }
    shared_data[local_idx] = value;

    workgroupBarrier();

    // Reduction in shared memory
    var stride = params.workgroup_size / 2u;
    while (stride > 0u) {
        if (local_idx < stride) {
            shared_data[local_idx] = max(shared_data[local_idx], shared_data[local_idx + stride]);
        }
        workgroupBarrier();
        stride = stride / 2u;
    }

    // Write the result to the output buffer
    if (local_idx == 0u) {
        output[workgroup_id.x] = shared_data[0];
    }
}

@compute @workgroup_size(256)
fn min(@builtin(global_invocation_id) global_id: vec3<u32>, @builtin(local_invocation_id) local_id: vec3<u32>, @builtin(workgroup_id) workgroup_id: vec3<u32>) {
    let global_idx = global_id.x;
    let local_idx = local_id.x;

    // Load data into shared memory
    var value: f64 = 1.7976931348623157e+308; // DBL_MAX
    if (global_idx < params.array_size) {
        value = input[global_idx];
    }
    shared_data[local_idx] = value;

    workgroupBarrier();

    // Reduction in shared memory
    var stride = params.workgroup_size / 2u;
    while (stride > 0u) {
        if (local_idx < stride) {
            shared_data[local_idx] = min(shared_data[local_idx], shared_data[local_idx + stride]);
        }
        workgroupBarrier();
        stride = stride / 2u;
    }

    // Write the result to the output buffer
    if (local_idx == 0u) {
        output[workgroup_id.x] = shared_data[0];
    }
}