Benchmarks are ran with different measurement times, to compare how error scales with measurement time.