Asus Skylake Z170 Motherboards: Maximus VIII Gene vs. Z170-A

Viewing page 5 of 7 pages. Previous 1 2 3 4 5 6 7 Next


Power Consumption

Note: Motherboards sometimes vary in regard to the BLCK and Turbo Boost settings with models. As higher clock speeds typically consuming more power, this makes direct power consumption comparisons unfair. For these tests, boards using the same CPU have been tweaked to use the exact same clock speeds, both CPU and GPU, to ensure a level playing field.

Compared to our Skylake review board, the Gigabyte Z170X-UD5, both Asus boards have noticeably higher power consumption under light load. Sitting idle, the Z170X-UD5 uses a substantial 6W and 9W less than the Z170-A and Maximus VIII Gene respectively. The difference shrinks somewhat during video playback.

During our heavy load tests, the Asus boards continue to draw more power, especially the Z170-A, possibly due to poorer power regulation. In real life tasks, the difference is moderate while a much larger discrepancy is observed with Prime95, the most demanding (though synthetic) stress test.

In terms of CPU and GPU performance, boards of the same chipset perform very closely to one another but these power differences made me wonder whether there was some cheating going on. A quick benchmark check revealed identical results in the video encoding tests, but a ~10% improvement for the Asus boards in Lost Planet 2, with the Maximus VIII Gene having a slight edge over the Z170-A. Believing this was an anomaly I also tried out our Crysis Demo benchmark and discovered a similar framerate increase.

This is puzzling as the CPU/RAM/GPU clocks and settings were identical in the BIOS and observed to be the same with monitoring software during testing. This puts into question the veracity of our original HD 530 GPU though I cannot say whether a typical Skylake board will perform better or worse without more data.

The TDP limit comes into play with all three boards when adding a graphical load (in the form of FurMark) to the Prime95 stress test as only a small increase in system power consumption is observed when doing so. Setting the CPU to 800 MHz (the lowest allowable without changing the BCLK) reveals the potential of the integrated graphics chip when untethered The difference for the three boards is about the same, between 30W and 32W, which contradicts the GPU performance anomaly detected earlier. If the graphics chip were secretly overclocked or receiving some other advantage, one would expect to see a more substantial power difference here.


To test the board's cooling, the CPU is stressed for ~15 minutes with Prime95 and FurMark. Temperatures of the boards' chipset heatsinks are recorded using a spot thermometer. The highest temperatures are taken for comparison.

The Z170X-UD5 and Maximus VIII Gene, which both happen to utilize VRM heatsinks connected via heatpipe, run noticeably cooler under full load than the Z170-A. That being said, the peak external VRM temperature seems proportional to the system power draw, which varies greatly between the three models. If you assume a linear relationship between power consumption and temperature, the Z170-A does appear to be at a disadvantage, though the plastic shroud covering half of the side heatsink may be equally at fault as the actual heatsink design.

Boot Performance

To test boot time, the BIOS/UEFI is optimized by setting the hard drive recognition and other delays to minimum, taking care not to disable common functionality like full USB support, POST messages, etc. and the time it takes to reach the Windows loading screen (it's stopped here because this is the point where the O/S and drive become factors) is measured.

Our trio of Skylake boards take an unusually long time to reach the POST screen, resulting in boot delays about double that of Haswell models. Of the three, the Gigabyte board is the quickest but only by 1~2 seconds.

Previous 1 2 3 4 5 6 7 Next

CPUs|Motherboards - Article Index
Help support this site, buy from one of our affiliate retailers!