yum-mirror/slang

Making it easier to work with shaders

git clone https://git.yummers.dev/yum-mirror/slang

Sai Praveen BangaruAdded diagnostics & built-in type lowering for `[CUDAKernel]` functions (#4042)52b91231c

master
branchesmasterall refs →
777 B22 linesraw
1//TEST:SIMPLE(filecheck=CUDA): -target cuda -line-directive-mode none
2//TEST:SIMPLE(filecheck=TORCH): -target torch -line-directive-mode none
3
4
5[AutoPyBindCUDA]
6[CUDAKernel]
7void plain_copy(float3[4] input, TensorView<float> output)
8{
9    // CUDA: __global__ void __kernel__plain_copy(FixedArray<_VectorStorage_float3_0, 4>  input_0, TensorView output_0)
10    // TORCH:    void __kernel__plain_copy(FixedArray<_VectorStorage_float3_0, 4>  _0, TensorView _1);
11
12    // Get the 'global' index of this thread.
13    uint3 dispatchIdx = cudaThreadIdx() + cudaBlockIdx() * cudaBlockDim();
14
15    // If the thread index is beyond the input size, exit early.
16    if (dispatchIdx.x >= 1)
17        return;
18
19    output[0] = input[0].x;
20    output[1] = input[2].y;
21    output[2] = input[3].z;
22}