However in computer systems, benchmarks are used as a method that tests the computer with a specifically designed program, and it takes measurements at the same points during every run. These points are then calculated to take an average. Based on the type of the benchmarking the result scale can vary.
A timed benchmark is one that runs a particular process and records the timings it takes to complete. A declined time would display raise in performance; as a faster computer can process and complete the benchmark in quick time, it provides the best performance in all parts.
In a benchmark where the computer is set to process something in a specific time frame, the result would be calculated by the number of times the computer processes the data again and again in that timeframe the benchmark runs. In this type of benchmark, a higher result would show hiked performance, as the computer would be processing more data in the similar frame of time.