Batman: Arkham City
To warm up this latest batch of Radeons, we grappled and glided our way around Gotham, occasionally touching down to mingle with the inhabitants.
We tested at a 1080p resolution with DirectX 11 effects enabled, normal tessellation, high FXAA antialiasing, and everything else cranked up. Those settings seemed to yield the best compromise of image quality and smoothness on the 7870.
We should preface the results with a little primer on our testing methodology. Along with measuring average frames per second, we delve inside the second to look at frame rendering times. Studying the time taken to render each frame gives us a better sense of playability, because it highlights issues like skipping, stuttering, and microstuttering that can all occur—and be felt by the player—within the span of one second. Charting frame times shows these issues clear as day, while charting average frames per second obscures them.
For example, imagine one hypothetical second of gameplay. Almost all frames in that second are rendered in 16.7 ms, but the game briefly hangs, taking a disproportionate 100 ms to produce one frame and then catching up by cranking out the next frame in 5 ms—not an uncommon scenario. You're going to feel the game hitch, but the FPS counter will only report a dip from 60 to 56 FPS, which would suggest a negligible, imperceptible change. Looking inside the second helps us detect such skips, as well as other issues that conventional frame rate data measured in FPS tends to obscure.
We're going to start by charting frame times over the totality of a representative run for each card. That should give us an at-a-glance impression of overall playability, warts and all. (Note that, since we're looking at frame latencies, plots sitting lower on the Y axis indicate quicker cards.)
For some reason, Batman: Arkham City exhibits high frame latency spikes rather frequently. We recorded similar frame time spikes on faster cards tested with a faster CPU and quicker storage, as well as on slower cards running the game with its DirectX 11 mode disabled. We can conclude with reasonable certainty that the skipping is inherent to this game.
A cursory overview of the graphs above tells us the Radeon HD 7870 and 7850 are close, and perhaps a little faster overall, than their competitors and predecessors. It would also appear the GeForce GTX 560 Ti exhibits more latency spikes than its rivals from the AMD camp.
We can slice and dice our raw frame-time data in three ways, as you're about to see. By the way, we should caution that none of the graphs below can be construed as self-contained scoreboards; instead, they each show a different facet of the cards' performance, and they should be viewed as a whole along with the raw frame-time plots above.
Our first graph shows average frames per second. Though this metric doesn't account for irregularities in frame latencies, it does give us some sense of typical performance.
In our second graph, we're demarcating the threshold below which 99% of frames are rendered. The lower the threshold, the more fluid the game. This metric offers a sense of overall frame latency, but it filters out fringe cases.
Our last graph tells us how long each card worked on frames that took longer than 50 ms to render. Ideally, the result should be "0" for every card. That's because the illusion of smooth motion is likely to begin breaking down once frame latencies rise into that territory. (For reference, 50 ms frame times would work out to a 20 FPS average if they were sustained for a whole second.) In a nutshell, this metric tells us how badly each card skips during gameplay, if it does at all.
Now, let's put it all together.
Clearly, the Radeon HD 7870 GHz Edition is the fastest card of the pack. It has the highest average frame rates, the lowest 99th-percentile threshold, and the least dire frame latency spikes. The Radeon HD 7850 is also positioned rather favorably compared to our GeForce GTX 560 Ti, though it's no better than its most direct predecessor, the Radeon HD 6950, when we look at frame latency thresholds and worst-case scenarios. The difference between the 7850 and the 6950 is difficult to notice in this test.