The Truth About PC Power Consumption

SYSmark 2007 Preview

The SYSmark test was rather easy, as its runtime is rather long (80 minutes and up) and simply tracking the power requirement during the benchmark run resulted in an interesting diagram. Let's start with average power, and the power required to complete this benchmark:

SYSmark is close to reality when it comes to performing typical application scenarios. The difference between 98 W average power requirement for the Intel system and 124 W for the AMD system equals 26%. This means that, on average, the AMD system required 26% more power for the workload to complete. While the average power and power required over the benchmark runtime was similar in our first run with 3DMark06 and PCMark05 looping, the actual power required was much higher this time.

The difference in average power consumption was 26%, but the difference in total power required (Wh) to complete the entire benchmark is roughly 50%! How can this be? Let's look at the power requirement diagram to get the answer:

Click to enlarge

The power consumption diagram, which shows the power requirement for the entire duration of the SYSmark 2007 Preview run, makes very clear why the AMD system required 50% more energy to complete the benchmark, although the difference in average power requirement was only 26%. The Athlon 64 X2 5000+ system took 14 minutes longer to complete the benchmark! During this time, the Intel system was already running idle again.

  • trewinnard
    If you break it down to component level you can get an pretty accurate measurement of what your consumption should be. Checkout this PC power consumption breakdown| http://www.infobarrel.com/PC_power_consumption_explained

    Tbuzz
    Reply
  • trewinnard
    If you break it down to component level you can get an pretty accurate measurement of what your consumption should be. Checkout this PC power consumption breakdown| http://www.infobarrel.com/PC_power_consumption_explained

    Tbuzz
    Reply