And customers have longer memories that some might imagine.
#Amd radeon hd 3870 benchmark software#
The problem with all this strategy from NVIDIA? It never panned out: Quad-SLI was a flash in the pan and because of software incompatibility and upcoming Vista changes, Quad-SLI is essential a dead platform and users that bought it threw their money away. And, just as AMD is planning, the NVIDIA 7950 GX2 was used in pairs to offer Quad-SLI. It essentially had two 7900 GTX GPUs on it, scaled down in clock rates for power reasons, in a full-time SLI configuration and it worked in exactly the same fashion as the HD 3870 X2 we see here. The most prominent recent example, besides the mid-range HD 2600 X2 card, was NVIDIA’s GeForce 7950 GX2 card. This isn’t the first dual-GPU graphics card we have seen in recent years. The small micro-scale I have for weighing these things doesn’t go high enough for the HD 3870 X2 to register (seriously, it’s over the 16oz that is measures, but so is the 8800 Ultra) so you’ll have to just take my objective word for it. The heat is also well handled and nearly all of the hot air is quickly exhausted out the back of the case and not back into the chassis as many people complained about with the HD 3870 cooler design.Īnd yes, the new HD 3870 X2 is heavy for a single card – can’t get around that fact. The card never got louder than a single HD 3870 card and was definitely more quite than the single-slot coolers on the 8800 GT and the HD 3850 cards of this past fall. I touched on it briefly in the overview of the Asus Radeon HD 3870 X2 card, but the acoustics and heat dissipation of the new HD 3870 X2 card are impressive. Offering anywhere from 25-90% improvement benefits over AMD’s previous top performing card, a single Radeon HD 3870, the new 3870 X2 should find its way into a lot of enthusiast systems. When compared to AMD’s own previous single-card lineup, the HD 3870 X2 is the first high-end, enthusiast class performance part that the company has released since the Radeon X1950 XTX WAY back in September of 2006. First, the fact that it is essentially a pair of HD 3870 GPUs on a single card, yet can show substantial performance gains over two actual 3870s in CrossFire shows dedication in engineering. In terms of AMD-on-AMD action, the HD 3870 X2 is impressive for a couple of reasons.
WiC in particular showed abysmal performance across the board for AMD’s cards, not just the new HD 3870 X2. Call of Duty 4 was the other major title I might call a performance tie thanks to the shifting of lead from AMD to NVIDIA at the 2560×1600 settings.Ĭompany of Heroes, and especially World in Conflict, were the big losses for AMD. In the case of UT3, the performance quirk we saw will probably be fixed as see evident in the performance of the standard HD 3870 cards in CrossFire. Looking at Crysis and Unreal Tournament 3, the two newest games in our test suite, the performances between the 8800 Ultra and the AMD HD 3870 X2 were quite close. The performance gaps weren’t enormous by my measure but the HD 3870 X2 was definitely faster. The HD 3870 X2 was able to either best the NVIDIA 8800 Ultra at the top resolution of 2560×1600 that we tested or was able to win at even lower resolutions of 1920×1200 or so. Titles like Bioshock, Call of Juarez and Lost Planet were the wins for AMD in this case. The story isn’t as cut and dry as perhaps they would have liked, NVIDIA’s 8800 Ultra still wins its share of the benchmarks, but AMD is now closer than they have been in years to having a part on the same level as NVIDIA’s best. The scores are just a bit higher than the single 8800 Ultra though so the power consumption on the HD 3870 X2 isn’t completely out of whack with reality.ĪMD was adamant that the HD 3870 X2 was going to be competing for the performance crown against the best that NVIDIA had to offer and it would appear that AMD has done just that. Obviously the higher clocking of the GPUs has some affect on this.
At both idle and at load, the single HD 3870 X2 card used more power than our test setup with two standard cards in the system the results are actually quite surprising as we would have expected the more efficient use of power to be with the solution that used the fewest total components. It actually doesn’t too that badly, though the claims of the single HD 3870 X2 card using less power than a pair of HD 3870s appears to be incorrect.
Obviously putting two GPUs on a single graphics card is going to result in some larger than normal power usage, but how exactly does the new HD 3870 X2 fare?
0 kommentar(er)
