cuda::experimental::launch

Defined in include/cuda/experimental/__launch/launch.cuh

template<typename ...ExpArgs, typename ...ActArgs, typename ...Levels>
void cuda::experimental::launch(::cuda::stream_ref stream, const hierarchy_dimensions<Levels...> &dims, void (*kernel)(hierarchy_dimensions<Levels...>, ExpArgs...), ActArgs&&... args)

Launch a kernel function with specified thread hierarchy and arguments.

Launches a kernel function on the specified stream and with specified thread hierarchy. Kernel function is a function with global annotation. Function might or might not accept the hierarchy as its first argument.

Snippet

#include <cstdio>
#include <cuda/experimental/launch.cuh>

template <typename Dimensions>
__global__ void kernel(Dimensions dims, unsigned int thread_to_print) {
    if (dims.rank(cudax::thread, cudax::grid) == thread_to_print) {
        printf("Hello from the GPU\n");
    }
}

void launch_kernel(cuda::stream_ref stream) {
    auto dims = cudax::make_hierarchy(cudax::block_dims<128>(), cudax::grid_dims(4));

    cudax::launch(stream, dims, kernel<decltype(dims)>, 42);
}

Parameters

• stream – cuda::stream_ref to launch the kernel into

• dims – thread hierarchy dimensions for this launch

• kernel – kernel function to be launched

• args – arguments to be passed into the kernel function