Average Performance And Efficiency
Changes in the way PCMark 8 calculates drive performance (compared to PCMark 7) mean that we no longer see 1000% boosts from an SSD upgrade. Gains up to 91% more accurately reflect user experience. That's great news for us; we don't skew our combined performance data towards other benchmarks, as we have in the past.
Adding another bar for average performance in the chart above would have made it super long. But we can still see the outcome in our average efficiency chart below. The slowest configuration sets the 100% baseline for my calculations.
Since efficiency can’t exceed 100%, I subtract “1” from each result before charting it. My overclocked $2550 PC, for example, consumes 148% more energy, but produces only twice (100% more) the performance.
Yet again, slow and steady doesn't win any races in the System Builder Marathon. Don’s $1300 machine consumes 16% less energy, but provides over 46% more performance than Paul’s baseline, yielding an efficiency gain of over 75%. Even my three-way SLI-equipped, Sandy Bridge-E-based build starts out with a 2.5% efficiency advantage (though it ends up over 18% behind after overclocking).