The Board - A Closer Look
Between the cooler and the SATA ports, I found the dual BIOS chips, which are controlled in a slightly different way than what we saw recently on the Z77X-UP5 TH. Instead of a switch on the board's edge, there are three switches on the rear I/O that control which BIOS is in use. There is also a Clear CMOS button, and a button labeled "O.C.", which immediately enables a preset overclock, even while in the OS. The blue and green switch, which swaps what BIOS is in use, lights up either the blue or green side according to the BIOS you are using with the board automatically enabling the backup BIOS should a boot failure be repeated a specific number of times.
The power switch is in the right corner of the board, while I found a blue RESET button down along the bottom right edge.
For monitoring duties, I found an ITE8728F, a very familiar Super I/O option that also provides the mouse/keyboard PS/2 port. The CPU VRM is controlled by an International Rectifier part labeled IR3570.
That of course manages the nine IR3550 powerstages that make up the CPU's VRM, each of which is capable of providing 60 A. I flipped the board over to find that there was nothing but some surface-mounted capacitors here, because those IR3550 parts integrate all the other parts you'd normally find behind the chokes, including Hi/Low MOSFETs, and the input driver. The design that puts all these parts together also allows for lower operating temperatures and greater efficiency, a big part of why Gigabyte has "certified" the Ultra Durable 5 parts, which carry these IR3550 powerstages, as "watercooling and K-SKU ready".
I found the VTT phase in the usual place, powered by an IR3553M powerstage, while chipset power is comprised of what appears to be two more traditional phases, although the design does seem a bit different from what I had expected.
The bank of DIMMs on the left side of the Gigabyte X79S-P5 WiFi is powered by yet another IR powerstage, this time an IR3570S, which in turn mates with another IR3553M, same as the VTT power phase. The IR3553 is rated for 40 A, which is more than enough.
The right bank is powered just as the left, using the same parts in all positions.
I found two different USB 3.0 controllers on the X79S-UP5 WiFi, one Fresno controller for the front ports, and a VLI part for the four ports on the rear.
There's a Marvell controller near the rear I/O to manage the eSATA ports, which is, most likely, a Marvell 88SE9172 part. The marking on the chip proved to be illegible, so I'm not 100% sure on that. The FireWire support is provided by a VIA part, a VT6308P shown in the second image above.
The dual LAN ports, like the USB 3.0 ports, are provided using different controllers, one an Intel part, and the other a Realtek part. Both are pretty reliable in my books, although some users do prefer the Intel part, so it's great to see that both types of users are covered here.
The audio on the board is provided by a high-end Realtek ALC898 CODEC, one that has proven itself in the field, but one that has, in my own testing, shown to be sensitive to circuit design. To swap the eight PCIe lanes from the lower x16 slot up to the x8 slot, I found four NXPL0408 parts, each one responsible for the switching of two lanes.
I also found a couple of PCIe clock buffers, both of which are made by ICS. Both of these parts were used on the X79 G1.Assassin2.
With the cooler removed from the board, you can see how the separate elements are organized. There are two large elements which connect to the PCH cooler with heatpipes but only one of them contacts the board, while the other sits on the VRM section for the left DIMM bank. Flipping the cooler over, we see contact pads on the VRM section that mates with the IR3550 powerstages. The other clearly has large standoffs that keep it from contacting the board's surface. The PCH contact plate only contacts the heatpipes on one side, which seems a bit unbalanced, but it seemed obvious to me that this was due to space restrictions after I had the cooler mounted back on the board.