Applications vary wildly in what they demand from a system, making it difficult for a single benchmark to provide a broadly representative score. Benchmark suites try to address this by running a set of workloads that hopefully capture a range of typical…