That chart compares theoretical peak GFLOPS per watt, which is an absolutely terrible metric for server workloads. As the paper I linked to showed, Atom is some

That chart compares theoretical peak GFLOPS per watt, which is an absolutely terrible metric for server workloads. As the paper I linked to showed, Atom is sometimes much better and sometimes much worse than large, fast server processors in performance per watt on real server workloads. Those charts don't take into account things like cache, or branch prediction, or multicore memory coherence. The only way to get meaningful performance per watt data is to actually run real workloads on various processors.