runmat-accelerate 0.4.1

Pluggable GPU acceleration layer for RunMat (CUDA, ROCm, Metal, Vulkan/Spir-V)
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

use std::fmt::Write;

pub fn build_sub2ind_shader(
    scalar_ty: &str,
    dims: &[u32],
    strides: &[u32],
    scalar_mask: &[u32],
    workgroup_size: u32,
    epsilon: &str,
) -> String {
    assert_eq!(dims.len(), strides.len());
    assert_eq!(dims.len(), scalar_mask.len());

    let mut shader = String::new();
    writeln!(shader, "struct Tensor {{ data: array<{scalar_ty}>, }};").unwrap();
    writeln!(
        shader,
        "struct ErrorState {{ code: u32, dim: u32, extra: u32, pad: u32 }};"
    )
    .unwrap();
    writeln!(shader, "struct Params {{ len: u32, }}").unwrap();
    for (idx, _) in dims.iter().enumerate() {
        writeln!(
            shader,
            "@group(0) @binding({}) var<storage, read> input{}: Tensor;",
            idx, idx
        )
        .unwrap();
    }
    writeln!(
        shader,
        "@group(0) @binding({}) var<storage, read_write> output: Tensor;",
        dims.len()
    )
    .unwrap();
    writeln!(
        shader,
        "@group(0) @binding({}) var<storage, read_write> error: ErrorState;",
        dims.len() + 1
    )
    .unwrap();
    writeln!(
        shader,
        "@group(0) @binding({}) var<uniform> params: Params;",
        dims.len() + 2
    )
    .unwrap();
    writeln!(shader, "const EPSILON: {scalar_ty} = {epsilon};").unwrap();
    let max_val = if scalar_ty == "f32" {
        "3.4028234663852886e38"
    } else {
        "1.7976931348623157e308"
    };
    writeln!(
        shader,
        "fn isfinite_scalar(x: {scalar_ty}) -> bool {{ return (x == x) && (abs(x) < {scalar_ty}({max_val})); }}"
    )
    .unwrap();
    for (idx, dim) in dims.iter().enumerate() {
        writeln!(shader, "const DIM_{idx}: u32 = {dim}u;").unwrap();
    }
    for (idx, stride) in strides.iter().enumerate() {
        writeln!(shader, "const STRIDE_{idx}: u32 = {stride}u;").unwrap();
    }
    for (idx, mask) in scalar_mask.iter().enumerate() {
        writeln!(shader, "const SCALAR_MASK_{idx}: u32 = {mask}u;").unwrap();
    }
    writeln!(
        shader,
        "fn set_error(code: u32, dim: u32, extra: u32) {{
    if (error.code != 0u) {{ return; }}
    error.code = code;
    error.dim = dim;
    error.extra = extra;
}}"
    )
    .unwrap();
    writeln!(
        shader,
        "@compute @workgroup_size({workgroup_size}, 1, 1)
fn main(@builtin(global_invocation_id) gid: vec3<u32>) {{
    let idx = gid.x;
    if (idx >= params.len) {{
        return;
    }}
    if (error.code != 0u) {{
        return;
    }}
    var offset: u32 = 0u;"
    )
    .unwrap();

    for (idx, _) in dims.iter().enumerate() {
        let dim_id = (idx + 1) as u32;
        writeln!(
            shader,
            "    {{
        var raw: {scalar_ty};
        if (SCALAR_MASK_{idx} != 0u) {{
            raw = input{idx}.data[0u];
        }} else {{
            raw = input{idx}.data[idx];
        }}
        if (!isfinite_scalar(raw)) {{
            set_error(1u, {dim_id}u, 0u);
            return;
        }}
        let rounded: {scalar_ty} = round(raw);
        if (abs(rounded - raw) > EPSILON) {{
            set_error(2u, {dim_id}u, 0u);
            return;
        }}
        let int_val: i32 = i32(rounded);
        if (int_val < 1 || int_val > i32(DIM_{idx})) {{
            set_error(3u, {dim_id}u, u32(int_val));
            return;
        }}
        let term: u32 = (u32(int_val) - 1u) * STRIDE_{idx};
        offset = offset + term;
    }}"
        )
        .unwrap();
    }

    writeln!(
        shader,
        "    output.data[idx] = {scalar_ty}(offset + 1u);
}}"
    )
    .unwrap();

    shader
}