Gigabyte X79-UD7 - The Overclocking Motherboard That Never Was

Power Consumption and Temperatures


We wanted to see whether or not the Gigabyte X79 UD7 board delivered more power to the CPU. We also wanted to observe the board’s power usage both under load and idle. Based on our tests, it is evident that the UD7 board does draw more power than the Intel DX79SI board. However, this is because the UD7 delivers more power to the CPU, specifically, 6 more watts to the CPU cores directly. Interestingly enough, the difference between the boards’ power consumption when it came to the CPU was only 1w, yet the CPU cores gained 6w in power on the UD7. To us, this means that the Gigabyte board is more capable of delivering more power to the CPU cores using nearly the same amount of total power.

In terms of temperatures, we did not notice a significant difference between the CPU core temperatures. Unfortunately, during our testing, the Intel board would not report its temperatures to AIDA64, so we were unable to compare the motherboard temperatures against each other. In general, both boards were relatively quiet, cool, and did not cause us any overheating issues. There was a recent occurence of someone taking a UD3 motherboard to its absolute maximum capability and somehow burned out a MOSFET. This type of occurrence is extremely rare and we are surprised that it happened at all, considering the fact that the BIOS has so many thermal and current protections built in. If anything, the occurrence of someone burning their board was a combination of poor cooling (he wasn't running any fans over the heatsinks) and a BIOS that failed to detect an overheating MOSFET. In this case, it was quite a catastrophic failure which was surprisingly all caught on video.

Overclocking

Cinebench at 4.6GHz (note 4.5G score is below that and the stock 10.41 below that)

We primarily focused on two aspects of the board's overclocking, the CPU and the memory. We found that the Intel CPU does not function well when both the memory and CPU are pushed to their respective maximums, so the user is only able to really push the ability of one or the other. We were a little disappointed that we could not get the CPU to remain stable above an overclock of 4.6GHz regardless of the voltage we poured into it. We also attempted the same settings that we had run on the Intel board in our 3960X review, but could not get those to remain stable either. Interestingly enough, when we tried those settings, upon entering the OS we noticed that CPU-Z was giving us a slightly lower clockspeed. Upon review, we noticed that the multiplier had magically gone down by one and that it was now sitting at 4.7GHz rather than 4.8GHz due to the 36x instead of 37x multiplier. For 4.6GHz, we set the BIOS voltage for the CPU cores to 1.43v, which is reasonable, but seemed to be the maximum we could use and still enter the OS.

When it came to memory overclocking, this board did quite a good job. We were able to review multiple kits of RAM from Patriot, Kingston and Corsair and reach very stable overclocks. The best OC we achieved was using two 1600MHz kits and clocking them up to 2133MHz. This overclocked set of RAM actually outperformed the 2133MHz kit that we have in our lab. Furthermore, we were able to overclock the 2133MHz kit we have to 2400MHz+. We also have a 2400MHz kit that we tried to review, but up until recently the system simply refused to function with a 2400MHz kit installed. Once the F7 BIOS came out, the 2400MHz kit of Kingston HyperX RAM worked perfectly. With F8, it is just as stable, perhaps even more so. We did try overclocking the 2400MHz kit, but without any success. It must be kept in mind that 2400 MHz is already considered a fairly substantial overclock for the X79 chipset.