In the previous post I mention the following announcements in my kernel ring buffer (visible through the dmesg utility):
[Firmware Bug]: AMD-Vi: IOAPIC[0] not in IVRS table
[Firmware Bug]: AMD-Vi: No southbridge IOAPIC found
[Firmware Bug]: AMD-Vi: Disabling interrupt remapping
I was running BIOS version f5 and found that the bugs above seemed to be corrected by updating the GA-F2A88XM-D3H to the newest beta firmware (f6b). Don’t worry about it being a beta, Gigabyte assures its customers that beta BIOS firmware is reliable, but that it simply may not include all of the features to be eventually included in the final release.
Unfortunately, despite the bugs’ apparent absence from my first reboot after installing the BIOS update, subsequent reboots throw the error as usual. I’ll describe the BIOS update below nonetheless:
The BIOS file is downloaded from Gigabyte’s website (http://www.gigabyte.us/products/product-page.aspx?pid=4716#ov, in this case) as a Windows executable named mb_bios_ga-f2a88xm-d3h_f6b.exe. It’s a self-extracting executable, and I unpacked it using 7zip via wine on my Linux machine (a nice use for wine). From KDE, I simply right-clicked the file in Dolphin and chose “Open with wine Windows program loader” – worked like a charm.
The BIOS file itself is 8 MB in size and named “A88XMD3H.F6b”. You simply copy it to a USB drive, insert the drive into a port and reboot the machine into the BIOS. Enter the Q-Flash Utility, select the file, and away you go.
In the BIOS release notes on Gigabyte’s website (see above link), the following improvements are listed:
- Beta BIOS
- Update APU AGESA code
- Improve APU OC performance
- Modified Dual-link DVI compatbility
- Add cTDP function
Number two (2) refers to the AMD Generic Encapsulated Software Architecture, which is a bootstrapping protocol that allows for the initialization of system components on the motherboard.
Number three (3) is interesting; I’ll try to run some benchmarks later to see if the overclocking performance has actually improved at all with my A10-6800k APU.
Number four (4) likely has no bearing on me, since I haven’t had any issues with my DVI output.
Number five (5) refers to Configurable Thermal Design Power and is an Intel technology, so I’m not concerned with that.
Engineering Walden 
I know this blog entry is more than a year old, but saying that
“Number five (5) refers to Configurable Thermal Design Power and is an Intel technology, so I’m not concerned with that.”
is incorrect. cTDP is a feature of the then-new Kaveri core (APUs starting with 7xxx) to limit its own power dissipation. Entirely optional, of course, some boards allow you to pick any value between 45W and 65W, but most of them just allow these two choices.
Hey, I always appreciate correction. Thanks for pointing out the error! I don’t know why I thought that a year ago, but you are clearly correct that support for cTDP is included in both AMD and Intel processors, and it is not an Intel technology (which wouldn’t make sense to include in an AMD motherboard, anyway, so I really don’t know what I was thinking).