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

// Auto-generated arange operation for u32

const WORKGROUP_SIZE: u32 = 256u;

struct ArangeParams {
    numel: u32,
    start: f32,
    step: f32,
}

@group(0) @binding(0) var<storage, read_write> arange_out: array<u32>;
@group(0) @binding(1) var<uniform> arange_params: ArangeParams;

@compute @workgroup_size(256)
fn arange_u32(@builtin(global_invocation_id) gid: vec3<u32>) {
    let idx = gid.x;
    if (idx < arange_params.numel) {
        let value = arange_params.start + arange_params.step * f32(idx);
        arange_out[idx] = u32(value);
    }
}