"Conclusion First of all lets talk about prices. Let's consider that the two R9 280X together cost 422 Euro. Compare that to the 396 Euro you'd have to pay for the cheapest reference R9 290X we can find. Therefore this one high-end card is is about 7 percent cheaper than the two mid-range models. To dive a bit deeper into the results, we start with performance differences in 3DMark graphics score where we see that the CrossFire of the two R9 280X is between 39 and 41 percent faster than the single R9 290X. In the next theoretical test we ran, Unigine Heaven, we see that the two R9 280X's are on average 17 slower than the R9 290X. In case of games it turns out that for example in Sleeping Dogs one the one card is about 18 percent quicker than the two cards in CrossFire. A closer look at Call of Duty Black Ops 2 shows the other end of the scale, where the two R9 280X's score 24 percent better than the single R9 290X. In this case we see a rather huge variance in the test results. Overall the two R9 280X cards are performing really well and there is no doubt the two of them are seriously quick, actually, on average they're 10 pecent quicker than one R9 290X. But the problem is that the average value is not even telling half the story. It turns out that in quite a few games, the two 280X cards together are slower than the one R9 290X, which make performance rather unpredictable. Other than that there is power consumption . In idle the power consumption of our test system is almost 70 percent higher with the two R9 280X then with one R9 290X and under load conditions the setup in CrossFire can almost be used as a heating system. Concluding, it can be said, that under certain circumstances the two R9 280X cards are quicker than one R9 290X but if you want to invest your money in an as sustainable way as possible, then we'd recommend you to buy a single card and even save a few bucks. " 基本属于浪费感情和电费