While you may be aware that the number of VRM phases does NOT in itself determine how much "power" is available to the CPU, many readers may not. More phases can reduce voltage ripple which "might" but won't automatically improve overclocking capabilities - as your test clearly demonstrated and has been demonstrated over and over in other reviews.
Gigabyte has had VRM design issues for years on both AMD and Intel mobos despite the claims that their premium priced mobos are "designed for overclockers". Both Intel and AMD mobos have demonstrated overheating issues with high frequencies and/or AMD FX-8000 series CPUs, which draw a lot of power at 4.4 GHz. and higher.
This 32-phase design looks like a typical knee-jerk reaction to the VRM overheating issue. Gigabyte seems to get half of the solution sorted but misses the other half. While they kept the VRM circuit below the 90C threashold where it starts throttling the CPU frequency, it did this at the expense of a 25% higher power consumption at idle. That's what we refer to in the Biz as "PPE" or Piss Poor Engineering.
Gigabyte has also responded in a half-arsed way to consumers desire for 2,500 mobo fan headers. We all know that every overclocker needs at least 2,499 fans for best performance so every mobo should have a minimum of 2,500 just to be safe... As with some other mobo makers such as Asrock, not all of the mobo fan headers properly control the fan speed be it 4 or 3 pin fans/headers. Controlling the fan speed is either via voltage or PWM. This ain't rocker science so one has to wonder why the mobo makers can't deal with such a simple electical concept?
Because many consumers have minimal electrical/electronic expertise they can easily be duped by meritless performance claims for mobos and other PC hardware. A perfect example is VRM phases. As noted more phases does NOT automatically equate to more "power" availability. In addition no one to date has ever shown that a 8-12-14 or 32 phase VRM mobo overclocks the CPU any higher than a properly designed 4-phase VRM circuit mobo. Typically more phases can result in a lower VRM circuit temp but as Gigabyte proves lower temp does not mean lower power consumption in the case of their 32-phase VRM.
In regards to one mobo overclocking better than another, often the difference is in the BIOS setting - things that we don't get to change - for our own good, so as not to FUBAR the mobo/CPU forever. Asus likes to push the timing envelop and it makes their mobos very inconsistent as does the fact their mobos are made by several companies for them.
I personally would not pay the signnificant price premium being asked for the Giga Z77X-UP7, Asus Z77-V premium or Asrock Z77/Z79 Extreme 11 boards as these tend to be fleecing of the naive. As your test shows you could not even reach 4.6 GHz. on the Giga board yet you can on other basic mobo models. These pemium priced mobos are for true workstation use or for people with more money than good judgment.
Anyone would be hard pressed to buy any of the afore mentioned mobos for better overclocking capability as none of them are better than much more reasonably priced models.