Gigabyte A75M-D2H and A75-UD4H Motherboard Reviewformerstaff -
» Discuss this article (10)
One of the first things I took notice of when I started having a look at these boards is that AMD makes a large point that these are dual-PCIE boards and that you can run AMD Dual Graphics and CrossFire on them. This leads to an interesting question: What portion of the gaming public is going to purchase a motherboard hovering around $100 and a CPU that is in a similar range, and then turn around and put three times that amount in graphics cards inside? I am not sure I know the answer to that question, but I felt obligated to check out that scenario.
The first option to adding graphics capacity to an FM1-based system, such as the Gigabyte A75M-D2H or A75-UD4H, is to make use of the AMD Dual Graphics feature. With this option, you can add a discreet graphics cards of the same family as the onboard graphics and they will work in tandem with each other, in effect doubling your graphics power.
The second option for increasing your graphics power is to override the onboard graphics by installing a discreet graphics card or taking advantage of the CrossFireX capabilities. I put a pair of XFX Black Edition 7970s in the GA-A75-UD4H and fired up a few of the graphics tests we use in our normal benchmarking lineup to see if, and how well, an AMD APU and the Gigabyte UD4H would make use with all that graphics horsepower at its disposal.
The first test was 3DMark 11 and I was very surprised at the result. I also tracked the relation between the CPU activity and load in relation to the load or capacity of the GPUs being used. The score of P9717 is almost exactly on par with a GTX 590 and 1200 points higher than a single XFX Black Edition 7970.
Next I ran the AvP benchmark in consecutive runs with ECC overriding AA to x16. The result is an impressive 133 FPS. Remember, this system consists of a $110 motherboard paired with an APU.
For the last look into CrossFire on the Gigabyte GA-A75-UD4H, I used FRAPS to monitor a scene in BF:BC2 at high settings and 4xAA. The result was an average of 148 FPS. BF:BC2 will use as many cores/threads as you have to throw at it, and you can see that the CPU and GPU activity is a bit different here, though no less impressive.
There are a few ways to parse this performance outcome. Given the low cost of pairing these boards with a low cost APU, it does make for some interesting possibilities for those who spend half their time cruising the net and the other half as a demanding gamer. Time to draw a conclusion on this pair of FM1 boards.