max_kernel.cu

__global__ void maxKernel(
    const int n,
    const float eps,
    const float* x,
    float* y)
{
    for (int i = blockIdx.x * blockDim.x + threadIdx.x;
         i < n; i += gridDim.x * blockDim.x)
    {
        y[i] = fmaxf(x[i], eps);
    }
}