Gigabyte X99-UD4P Haswell-E Motherboard

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

TEST RESULTS

Power Consumption

As the X99-UD4P is the first X99 LGA2011-v3 board we've examined, I will compare it to a variety of Z97 LGA1150 models. A direct apples-to-apples comparison isn't possible as the CPU is different, the board requires a discrete graphics card, and most users choosing this platform will use at least four DIMMs to take advantage of the quad-channel memory support. To get a better sense of the relative power differences between the two platforms, I've included results for one Z97 model, the Gigabyte Z97MX-Gaming 5, using the same video card and memory configurations. All results are with 2 DIMMs unless otherwise noted.

Note: Motherboards vary in regard to stock Turbo Boost settings so more aggressive models consume more power which is a detriment when using energy efficiency as a metric. For these tests, boards using the same CPU have been tweaked to use the exact same clock speeds (e.g. a multiplier of 39x/38x/37x/36x for 1/2/3/4 core operation for the Core i7-4770K and an integrated GPU frequency of 1250 MHz) to ensure a level playing field.

On light load, adding the low-end Radeon HD 5450 to the i7-4770K/Z97MX-G5 combination adds 9~10W to its total power draw. If you subtract that from the i7-5960X/X99-UD4P combination, its idle consumption is similar to the Z97 models compared. Video playback however, is substantially more demanding even though it shouldn't use that many CPU cycles regardless of the graphics solution.

On the X99 board, have two extra DIMMs in this spot doesn't make too much of a difference, and in fact the power draw is actually slightly lower for some reason when rendering Flash video. On the Z97MX-G5, moving to four sticks of memory has a negligible effect as well, though it utilizes only dual channel operation just as it does with two sticks.

The beastly i7-5960X's 140W TDP is quite evident when the system is placed on heavy load. After accounting for the video card, video encoding draws about 35W extra compared to i7-4770K/Z97 combinations and that figure rises to more than 50W for Prime95. Doubling the memory has a moderate affect on the Z97MX-G5, increasing power consumption by 5W and 11W in these two tests respectively, while for the X99-UD4P, Prime95 works the quad channel memory hard enough to draw 33W more.

Cooling

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

Note: The Scythe Kabuto is the reference heatsink for our motherboard cooling tests but as it's not compatible with the LGA2011 socket, I substituted a Noctua NH-L12. The NH-L12 should be superior cooler for components around the socket but also keep in mind the i7-5960X is a much hotter chip running a more demanding memory subsystem, so the relatively temperature differences shouldn't be too far off.

Despite having quad-channel memory and a much hotter processor, the X99-UD4P's onboard cooling fares well. While the VRM heatsink is relatively small, being connected to the expansive chipset heatsink via a heatpipe seems to help. On full CPU load, the highest temperatures recorded on the heatsink exteriors are less than 20°C above the ambient temperature, giving it a slight edge over the past few Z97 models I've tested.

SATA 6 Gbps Performance

The X99-UD4P's "SATA3" connectors, i.e. the grey block and upper black ports, benchmarked noticeably higher using small 4K blocks than the non-RAID "sSATA3" ports, which is interesting as they're both hooked up to the same native Intel controller (which hasn't been changed in a few years AFAIK) and thus should provide the same speed. Overall, the faster ports perform more or less the same as older series 7, 8, and 9 chipset models.

Boot Performance

To test boot time, the BIOS/UEFI was 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 measured 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).

From a usability standpoint, booting up is the one area that I notice a dramatic difference compared to most Haswell systems. The X99-UD4P takes more than double the time to complete the POST sequence before launching the O/S. It does have a lot of devices to detect but not significantly more than Z97 solutions. Having to check an extra two sticks of memory makes the wait slightly longer still.



Previous 1 2 3 4 5 6 7 Next

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