Unfortunately, those stunning performance gains aren’t seen with any of these cards. While the RX 460 gains a roughly 10-percent boost in theory, that only slightly stretches the frame rate in reality. Once again, DX12 does nothing for these cards’ modest graphical capabilities and limited memory capacities. (DX12 adores having more than 2GB of RAM to play with.)
Once again, the RX 460’s results fall in between the two Nvidia cards. None of them deliver great frame rates, to be honest, but the RX 460’s 32 fps in DX12 is definitely playable in a slow-paced real-time strategy game like this—albeit with a top-of-the-line processor, which no doubt affects results in a CPU-centric game like this. You’d probably want to drop details to even lower levels in a standard rig.
Test 7: Synthetic benchmarks
We also tested the RX 460 and its rivals using 3DMark’s highly respected DX11 Fire Strike synthetic benchmark, which runs at 1080p, as well as its brand-new Time Spy benchmark, which tests DirectX 12 performance at 2560x1440 resolution.
Like I said in the intro: None of these are barn burners. But as expected if you’ve been paying attention to the real-world performance results thus far, the RX 460 splits the difference between the two Nvidia cards.
Test 8: Power
We test power under load by plugging the entire system into a Watts Up meter, running the intensive Division benchmark at 4K resolution, and noting the peak power draw. Idle power is measured after sitting on the Windows desktop for three minutes with no extra programs or processes running.
AMD’s Polaris GPU grants Radeon graphics cards a huge step forward in power efficiency, but it still isn’t as lean as Nvidia’s chips. The overclock and extra power connector see XFX’s Radeon RX 460 sipping exactly as much power as the GTX 950, while the power connector-less GTX 750 Ti comes in a full 30W behind. That said, the total system power consumption under load with all of these cards installed stayed far below 200W at peak. That’s nothing to sneeze at.
Test 9: Heat
We test heat during the same intensive Division benchmark, by running SpeedFan in the background and noting the maximum GPU temperature once the run is over. All three of our test samples use custom coolers; there’s not a single reference design in the mix.
Sign up for Computerworld eNewsletters.