Welcome Stranger to OCC!Login | Register

Gigabyte 990FXA-UD7 Rev 1.1 Review

» Discuss this article (11)


Testing for the products in this review will consist of putting the components through the OCC benchmarks suite which uses both synthetic and real world applications to score their relative and finite performance. The gaming benckmarks will consist of both actual game-play and synthetic testing as well to determine if a product offers a performance advantage over another. A new installation of Windows 7 has been installed for testing purposes as well as the latest non-beta drivers for the XFX HD R7970 Black edition. For the testing all turbo,low power state and any CPU throttling features have been disabled so as not to have any effect on the performance or the possability of skewing the results.


Testing Setup:



  • Gigabyte GA-990-FXA-UD7  200Mhz x 25 = 5000MHz

The Gigabyte 990-FXA-UD7 inspires overclocking confidence. The large effective heat sinks on the VRM MOSFETs, the amount of copper used in the PCB tracers, and the 8+2 CPU phase power design gave me no concerns about elevating voltages to go after a very high overclock. For this review I decided to simply take advantage of the unlocked multiplier that all the Bulldozers have to obtain my overclock. As I covered earlier in the review, neither Touch BIOS or Easy Tune 6 seem to work very well or inspire confidence that what you set in them is what you will get for a result. Thus I performed all of my overclocking in the UD7's very mature, standard BIOS, which is the method I prefer. Overclocking with the UD7 could not be easier. The BIOS is laid out very intuitively and once I disabled all of the power saving features, it was simply raising the multiplier and/or voltage after every round of stability testing. 

Gigabyte motherboards have a strong history of being able to extract the maximum frequency out of processors, and it seems the UD7 is no exception.



Load Line Calibration:


I would like to take you on a brief aside here for a moment, if I may. As I mentioned I have owned both versions of the UD7, the 1.0 version and revision 1.1 we have here today. The difference between the two is that rev 1.1 has added LLC or Load Line Control. The difference in performance, overclock, heat, and energy has turned out to be rather large. I thought I would take a minute here to illustrate the effect that LLC can make to the performance enthusiast.

Below are two charts I grabbed from the OCCT program while running the stability test. The first is the 1.0 revision of the UD7 without LLC. You will notice the ongoing varying voltage that exists. This varying voltage is called ripple and is due to incomplete suppression of the alternating waveform within the power supply. What you cannot see on this chart because it apparently happened too fast for it to be caught when the chart was generated, is that the voltage was set to 1.525v and dropped almost a full .10v down to 1.43v and occasionally would briefly drop to 1.39v. This was the same CPU and eventually I was only able to get a stable 4700MHz at a voltage of 1.556v.

Now consider the second chart that was taken using the same CPU in the revision 1.1 UD7. Using the extreme setting in LLC we get voltage that is almost perfectly steady and the CPU needs a lower voltage setting at even higher frequencies for stability. Using the same processor in revision 1.1 of the UD7, I was able to (so far) obtain a P95 and bench stable 5.0GHz at 1.488v; the overclock I am using for this review. The LLC function is not limited only to the CPU voltage. Once set to one of six settings, all of the incoming voltages demonstrate a 'flat' stability from all voltage rails similar to the 12v reading shown in the bottom image. Each of the LLC settings on the UD7 have different characteristics to them. In this particular case the 'Extreme' setting bumps the voltage slightly upward when a heavy load is applied whereas the 'High' setting produces a steady flat line when load is applied. Conversely, the medium setting will slightly drop the voltage and then flatten out when load is applied. This is a great feature and using the six settings within it, you can find the one that fits your needs and performance preferences.




Maximum Overclock:

All motherboard and CPU combinations have been tested for stability at the frequencies listed. Benchmarks will be run at stock speed and the listed over-clocks to determine the performance of each at both speeds.




Scientific & Data:

  1. PCMark 7
  2. HD Tune 5.0
  3. AIDA 2.20
  4. Apophysis
  5. WinRaR
  6. Geekbench
  7. Bibble 5
  8. Office 2007
  9. POV-Ray
  10. ProShow Gold
  11. HandBrake 9.5
  12. SiSoft Sandra
  13. ScienceMark
  14. Cinebench


  1. DiRT 3
  2. Battlefield 3
  3. 3DMark 11


Related Products
Random Pic
© 2001-2018 Overclockers Club ® Privacy Policy
Elapsed: 0.3092100620   (xlweb1)