Power Consumption

As is usual now, Gigabyte includes its Dynamic Energy Saver software in the package of both boards - this adjusts the CPU power phase use on the fly according to load to maximise the efficiency of the hardware. The virtual twelve phase design jumps up in pairs from two to 12, although we found it doesn't go below four phases even when in its most efficient power-saving state.

Gigabyte does not adjust the base clock frequency at all, and as usual we found its design is limited to use in Windows only, unlike MSI's GreenPower which can be enabled and adjusted from the BIOS itself. In addition, the frequency of changes is better than Asus' EPU², but far less than MSI's GreenPower again.

With a Core i7 920 CPU, 6GB (3x2GB) of memory, a GeForce GTX 280 graphics card and the BIOS set to its default values we firstly tested with DES disabled, then enabled with the specific CPU throttling function on and off to see what difference, if any, it made.

The various power saving modes of the DS4 are in purple-violet, while the UD4P's are in red-orange.


Under its most efficient power saving mode, the DS4 is just a single watt more power hungry than the MSI Eclipse SLI, however while the MSI has somewhat more hardware on-board it also clocks the CPU back massively under its Max Power Saving option. The Gigabyte with its CPU Throttling enabled mirrors this by clocking the 2.6GHz Core i7 920 back to just 865MHz (6.5x133MHz), but according to the results above there's absolutely no point in doing this because when disabled at 1.5GHz it uses exactly the same amount of power. The DES factor saves 6W by modulating the power hardware on top of the the CPU alone in its underclocked C1E state.

In contrast the Gigabyte UD4P shows less difference between DES enabled and disabled than the Gigabyte DS4 - it's more power hungry in general, but it has more MOSFETs and a digital PCI-Express switch to power. The UD4P falls in around the MSI Eclipse SLI in terms of power requirements, but is better than the Asus by several watts.


Under load, with CPU throttling enabled the Gigabyte mirrors the MSI in clocking the CPU right down to just 1,500MHz, compared to the MSI's 1,425MHz, but it still uses 10W more power on both boards. We don't know why you'd buy a high end, performance quad core to then under clock it though, it seems a contradictory and pretty much useless feature on both boards. Is saving 45W really worth it when a 1.5GHz CPU is significantly slower to finish a task than a 2.6GHz one? The efficiency factor is completely void.

The MSI stretches a slight lead over the Gigabyte boards in the mid-table test where all the CPUs are running at full throttle, with only the motherboards' power regulation hardware under efficiency control, and the DS4 edges out a very slight power drop to the UD4P when the DES software has been disabled too.