| 412 | |
| 413 | template <typename F> |
| 414 | __global__ void parallel_for_kernel(const size_t n_elements, F fun) { |
| 415 | const size_t i = threadIdx.x + blockIdx.x * blockDim.x; |
| 416 | if (i >= n_elements) return; |
| 417 | |
| 418 | fun(i); |
| 419 | } |
| 420 | |
| 421 | template <typename F> |
| 422 | inline void parallel_for_gpu(uint32_t shmem_size, cudaStream_t stream, size_t n_elements, F&& fun) { |
nothing calls this directly
no outgoing calls
no test coverage detected