That means you can process more requests at once (across several threads/cores), but per request timings will be slower because each core is slower than the cores on the competitions CPUs.