I gate implementation using CUDA-Q.

// Compile: nvq++ main.cpp -o main
// Run:     ./main
 
#include <cudaq.h>
 
struct kernel {
    __qpu__ void operator()() {
        cudaq::qubit q;
        h(q);   // prepare |+>
        // identity is a no-op; the state is unchanged
        mz(q);
    }
};
 
int main() {
    auto counts = cudaq::sample(kernel{});
    counts.dump();  // ~50% |0>, ~50% |1>
}