Gigabyte 990FXA-UD7 Motherboard Review

Every once in awhile you come across a motherboard that raises the bar on what an enthusiast board should be. It sets the standard for features and performance exceeding all your expectations. This type of board forces every board you see from that point to meet a higher standard in order to impress you. The Gigabyte 990FXA-UD7 is not one of those boards.

continued...

Overclocking

Overclocking the 990FXA-UD7 wasn’t the best experience I’ve ever had nor was it the worst. I tried Gigabyte's Easy Boost overclocking feature which netted me a result of 3.2GHz after locking up once. This was nothing to write home about and I disregarded these settings in search of my own.

Keep in mind you would typically want to use a Black Edition CPU with a board like this which of course makes things much easier. However in order to separate the board from the CPU as best as we can, we really try and push the bus clocks on the board to showcase what non-Black Edition users can expect, as well as BE users who want to use a combination of multiplier adjustment and bus clocking to achieve their goals and really maximize performance. On this note alone, the 990FXA-UD7 wasn’t all that bad. It clocked to 245MHz without any adjustments aside from the CPU host clock. Given the myriad of problems I had with the board this was surprising to say the least. At 250MHz or so though I had to start reeling in the HTT frequency and try and keep it as close to that 2000MHz number as I could. Additionally because of the RAM clocking issues I had, I was constantly adjusting my memory dividers to keep the memory speeds as close to DDR3 1333MHz as possible. I was able to increase the CPU host clock to 260MHz but stability became a problem. I increased the NB voltage to 1.2v which corrected the issue and allowed me to continue.

I wasn’t able to get anymore out of the 990FXA-UD7 on the overclocking front. I’m not sure if the memory had anything to do with it, but I suspect that we might not really get a feel for what these 990FX boards are truly capable of until we see Zambezi parts in hand for testing. Below are the results obtained with my Phenom II 1055T test CPU.

3.64GHz (260.0 x14.0) DDR3 1386.8MHz

Article Image Article Image Article Image

Conclusion

Dan's Thoughts:

They say everyone makes mistakes. When I think of the Gigabyte 990FXA-UD7 I realize that truer words were never spoken. From start to finish my time with the 990FXA-UD7 was miserable. I just wanted the experience to end but with all the problems, working with the board took far longer than it should have. It is without a doubt one of the worst motherboards I’ve ever worked with. I had more problems with this board during my review than any other before it and I’ve been doing this for some time now. I’ve certainly not seen any boards this bad in the last two years. Ordinarily I’d say that Gigabyte has some work to do and that the problems with the 990FXA-UD7 can be worked out but that would be a serious understatement. I’d recommend sending it back to the drawing board. It’s just that bad. Well when I say bad it wasn’t as if the board wasn’t stable or anything like that. It just annoyed me with odd quirks time and time again often without explanation. I thought long and hard about how to word this but I’ll just let my testing notes paint the picture.

  • Legacy USB keyboard support is enabled by default, however it still doesn't work correctly. The USB keyboards do not work in the option ROMs or from a DOS boot environment such as the one I create Ghost images with. I tried three USB keyboards with USB to PS/2 adapters with no success. I had to drag out a crappy PS/2 keyboard with a busted numpad enter key on it to resolve the problem. (Really Gigabyte? Still struggling with this in 2011? )

  • System will not work with Corsair DDR3 2000MHz RAM at DDR3 1600MHz speeds. Attempt to run the memory at 1600MHz (timings on auto, and also tried on manual with voltage at 1.50v-1.65v) using the Corsair Dominator GT 2000MHz modules always resulted in a hard lockup at the POST screen above DDR3 1333MHz speeds. Tried restarting multiple times. Had to clear CMOS to get into BIOS and continue at one point. Switched to Corsair Vengeance DDR3 1600MHz 8GB (2x4GB) Kit. Corsair Vengeance kit did no better. Set timings manually on this kit as well and manually adjusted voltages. Also increased NB voltage, CPU voltages, and others slightly to see if I could stabilize the system without success. Having trouble getting past the POST screen with memory set to DDR3 1600MHz speeds on two different sets of RAM.

  • Got past the POST screen with Corsair Vengeance RAM running at DDR3 1600MHz speeds. I had adjusted timings down as loose as 11,11,11,28,1T @ 1.510v (which the system automatically does after a failed overclock anyway.) System BSODed before completing the OS load. OS partition totally corrupted. Restoring from Ghost image. This board simply won't handle DDR3 1600MHz speeds using any of the Corsair memory I have on hand. Unacceptable given the popularity of the Corsair Vengeance 8GB kit. Setting memory timings to automatic and selecting DDR3 1333MHz memory speed in order to continue with my testing. (This seems to be working fine. )

  • Consistent warm reboot issues. System will not perform restarts when restarted from a DOS environment, firmware / Option ROMs. It will only restart from Windows and not always 100% of the time. (WTH? I see cold boot issues all the time but warm boot problems? )

  • First attempt to install the OS resulted in the system showing an error writing to the drive. Rebooting the system and retrying allowed me to continue. OS installed slowly, but finished without anymore issues.

  • System will not boot from to the OS (If the OS is on a SB950 RAID array) if the Marvell 9172 controller for the eSATA ports is set to RAID mode. AHCI and legacy IDE modes work fine. The system boots fine with additional arrays on the SB950 controller, or if the Marvell 9172 (grey ports) are set to RAID mode, however the option ROM menu prompt is missing and I can't actually configure a RAID array to test them. I reinstalled the OS to a single AHCI volume and disabled RAID mode on the SB950 controller. Only then did the Marvell 9172 controllers work properly.

  • Windows installation is extremely slow. Appears to lock up at times with no response from the mouse cursor before continuing. Problem only exists when installing the OS to a RAID array. Problem didn't occur on a single AHCI volume.

  • While dual GTX 480's are installed, I do not see the POST screen or have any opportunity to enter the BIOS / Option ROM menus. After a restart from Windows it operates normally, however it sometimes hangs on the option ROMs.

  • After installing the second graphics card, Windows appeared to load the driver and then prompt for a restart. System hung on restart. After the restart the driver was not loaded and in fact showed a standard VGA adapter which wasn't working. I installed the drivers manually which seemed to take an unusually long time to do. After reinstalling I restarted the system and SLI worked properly afterwards. Problem verified on both my AHCI single drive install and my RAID 0 array install.

I’ve never wanted to run over a review board with my truck so bad in my entire life. I’d love to have shot it, burned it, or done something more spectacular to it in a fit of nerd rage. So far all the 990FX chipset based boards I’ve seen at this point have a long way to go before they are mature enough to be considered as viable upgrades or for use in new system builds. I’ve seen chipsets and boards with teething problems plenty of times before but these 990FX boards seem to be especially problematic. Therefore it may not be fair to lay the blame entirely on Gigabyte.

I don’t really understand what the problem is with the 990FX boards given that the chipset isn’t really anything we haven’t seen before and the last group of 890FX boards I worked with all ranged from good to excellent. As a result of my experiences I fear that it will take more than drivers and a BIOS fix to resolve the issues I encountered with the 990FXA-UD7. This is probably the worst motherboard I’ve ever reviewed here at the [H]. I can’t recommend it to anyone for any reason as it stands today and feel like it should be avoided at all costs. Obviously Gigabyte may be able to address some of these issues and even resolve them but I’m less than optimistic about that given just how bad this board truly was to work with.

Kyle's Thoughts:

Dan has beat on the board enough already, so I will not chime in any more except to say that he has some very valid points.

On the positive side, the board did load Windows 7 perfectly, although slowly. The BIOS updated perfectly. And the 990FXA-UD7 did pass our 24 hour incubated stock clock Torture Testing. I did get the board up to the 4GHz mark using a 1100T Black Edition processor at 1.38 vCore with 1600MHz RAM. 4.2GHz BSOD'd.

The Bottom Line

Gigabyte seemingly has put the 990FXA-UD7 motherboard on the market that has a lot of issues. You really need to read through the review to understand why. As it stands, there is no way we can suggest you purchase this motherboard. AMD needs to worried as well as this is the kind of product leading the charge for its late to the party Bulldozer/Zambezi.

Discussion