Gigabyte GA-EP45T-Extreme Motherboard Review

A Closer Look at the Gigabyte GA-EP45T-EXTREME

The image above highlights all the major players on this motherboard and the overall appearance of the board screams typical Gigabyte. This is definitely holding true so far with clashes of color all over the board that really make it possible for only a mother to love. Let's take a closer look now at the various areas of the motherboard and see if we can't make these colors look decent with selective framing of the images.

Like clockwork, we not only start in the same spot as always, but we will go around the board in a circular motion from the top of the board. Up here is the ever so important CPU socket, the heart of the beast if you will. Lining the rear and top edges of the 775pin LGA socket we find our "virtual" 12 Phase power station. The term virtual is quoted from the Gigabyte site but we couldn't find further explanation anywhere. There certainly aren't 12 physical phases so we will leave it at that. The low RDS(on) MOSFETs are cooled in part by a typical heat pipe setup that originates from the south bridge. This heat pipe configuration isn't exactly typical though, more on that later. As always we will be testing CPU cooler fitment with the Thermalright Ultra-120 for interference with said heat pipe roller coaster as well as the DIMM slots. At first glance though, it doesn't appear as if it will be an issue.

Moving over to the DIMM area we get another dose of the wonderful color scheme but more importantly, DIMM slots that look like they should play nice with big CPU coolers. The fact that memory slots are so far down the board these days means that CPU coolers can almost have free reign of the upper portion of the board. We also get all connections to the outside edge and a quick shot of the multi phase vDIMM circuit Gigabyte has designed. Like the CPU and what we will see with the NB, the vDIMM circuit is powered by the ferrite core chokes and low RDS MOSFETs.

Continuing our journey around the GA-EP45T-Extreme we find the major input/output area of the board down near the south bridge. Gigabyte has gone with a pair of SATA ports at a 90 degree angle with the last 4 being of a traditional orientation. Our dual BIOS, onboard LED diagnostic readouts, and onboard power/reset buttons are also found down here. We also have our IDE connector and controller, the ITE IT8213F. Front panel connectors as well as two USB onboard headers follow along the bottom of the motherboard and reach across to the rear of the board. Further along the bottom we have three headers for firewire connectors and one of four fan headers along with an onboard serial header and another ITE IC. The ITE IT8720F is a super I/O controller chip but interestingly enough, we can't find any information on it.

The expansion slot layout is very nice with a two slot gap between the two main PCI-E 16X slots and a PCI & PCI-E 1X slot above the top PCI-E 16X slot. This means, regardless of what GPUs you use, you add in soundcard guys can use a PCI slot no problem...except for length, but there should be enough room to the DIMM slots. The very top PCI-E 1X slot will have to be reserved for the smallest of small cards as there is no room for anything else. The very bottom PCI-E 16X slot is nice that it is full size, so any card will fit, but will only run at 4X. In the second photo the clock generator IC (ICS 9LPRS914EKLF) and battery are located underneath the heat pipe mess coming from the NB. We also have the Realtek ALC889A codec which isn't the highest quality on-board audio solution but should be good enough to get the job done. The Realtek RTL8111C gigabit network controller IC is also located right here.

We now turn our attention to the heat pipe cooling setup of this motherboard. There is no point in sugar coating this so here it is, this setup is an absolute waste of copper and board space. It looks fancy and the marketing department probably loves what they came up with, the only problem is that marketing shouldn't design a motherboard cooling solution.

First we are going to pull it off to have a look at it. Under where the heatsink was just sitting, we can see full coverage of thermal paste on the north bridge. The thermal paste used is that hard cruddy stuff which works fine, but we have found when a board flexes, it breaks its bond with the chipset and from there on out, the thermal transfer is hindered. This was the case with our board we believe as initial testing with the stock heat sink was fairly disappointing. We would like to see a soft thermal paste that doesn't dry right up and harden. The south bridge and PWM heat sinks use thermal pads to interface with the components and they seem to work just fine. Now, let's get into the finer details of why this heat sink/heat pipe setup is great, but had a major flaw.

We can see there is a single heat pipe going from the south bridge to the north bridge, so far, this is fine and the south bridge is adequately cooled. The north bridge, however is a different story all together.

Looking directly from the side, we can see a thin base plate that makes contact with the P45 NB, but sitting directly above that, is a series of cooling fins. These fins then attach to the nice and thick upper plate that has the water block, provides the base for the Hybrid Silent-Pipe Module, and finally the heat pipes that then carry on to the PWM areas.

Our question to Gigabyte is, how exactly do you think the water block, Hybrid Silent-Pipe Module, or those heat pipes are going to remove heat from the chipset? The cooling fins in between the two layers of the whole assembly are for dissipating heat, not transferring it. There is very little mass in contact with the chipset which is required to absorb large heat loads when the system is really being pushed. The argument here then is that the P45 chipset runs cool. Well, if it runs so cool, what is all this elaborate cooling for then? Is a water block even necessary? What about the heat pipes that are supposed to be transferring heat to the PWM area where the cooling fins then dissipate the heat? To put it bluntly, this design is probably the most inefficient we have seen to date. You can see the photos, come to your own conclusions.

I think it is high time motherboard manufacturers took a step back and started designing cooling solutions based on performance, not how it looks esthetically on a board because what we are looking at here with the GA-EP45T-Extreme is a great design ruined by what looks like a solution a marketing intern came up with.

Flipping the board over, we can see small back plates for the north bridge and south bridge heat sinks and these are a very welcome sight. They are sturdy and when combined with the screws/springs that hold the heat sinks in place, provide an excellent environment for good contact between the heat sink and the component. It is just too bad the quality design ends there. One other item we want to point out is the fact that we have additional MOSFETs on the back side of the motherboard around the CPU socket. Notice that these are not heat sinked in any way? Why is it that the topside MOSFETs need heat sinks with all those heat pipes? The answer: they don't. Those heat sinks are for dissipating the heat from the north bridge but as we just pointed out, there is no way for that to efficiently happen. In the end, this entire heat pipe assembly, water block, and Hybrid Silent-Pipe Module are nothing more than a marketing gimmick that provide very little, if any additional cooling in our opinion based on the design.

We apologize for the chastising of Gigabyte because they are not the only ones putting out these terrible cooling solutions, but this design is the creme of the crop and pushed us over the edge. Something needs to be said and users need to know that we are paying for elaborate cooling solutions that are a waste. A basic south bridge heatsink, and small all copper north bridge heat sink with a low RPM fan is all that is required with the P45 chipset. We know people hate the idea of a small fan in their system but if a little bit of effort was put into designing a good one that didn't die inside a month and scream the whole time, the roller coaster heat pipe alternative could be avoided and value of the motherboard could be increased.

For those that are interested, here are a couple photos of the multi phase north bridge and the "virtual 12-phase" CPU PWM circuit. Both are very well designed and in conjunction with the high quality MOSFETs, neither area gets even remotely warm under load. The finger test doesn't even register a temperature for either as it just isn't enough above ambient for our skin to tell the difference. Of course, this is with just a dual core. A quad core processor may warrant some sort of heat sink but a handful of Swiftech MC-14s would even be over kill, let alone a Thermalright solution. The stock heat sinks are again, more than enough to keep the CPU PWM cool as a cucumber as well.