Simple CUDA version of std::array

GenericArray.cuh

#include <cassert>

template<typename T, size_t N>
class GenericArray {
    T buffer[N];

 public:
    __host__ __device__ GenericArray() {
        memset(buffer, 0, N * sizeof(N));
    }
    __host__ __device__ GenericArray(const T *in) {
        memcpy(buffer, in, N * sizeof(N));
    }
    __host__ __device__ GenericArray(const GenericArray<T, N> &in) {
        memcpy(buffer, in.data(), N * sizeof(N));
    }
    __host__ __device__ GenericArray<T, N> &operator=(const GenericArray<T, N> &rhs) {
        if(this == &rhs)
            return *this;
        memcpy(buffer, rhs.data(), N * sizeof(N));
        return *this;
    }
    __host__ __device__ T &operator[](const size_t &index) {
        if (index > N) {
            assert(false);
            return buffer[0];
        }
        return buffer[index];
    }
    __host__ __device__ T *data() const {
        return (T *)&buffer[0];
    }
    __host__ __device__ constexpr size_t length() {
        return N;
    }
};