JMH benchmarks voor array reading because I want to find out the impact of if L1/L2 caches and prefetching on a java app.