Our first benchmark is a simple walk through the "Raffle Square" section, which lies at the beginning of the game, before everything goes south and combat is introduced. This area of the game has crowds, "god ray" effects, and a floating Benjamin Franklin facsimile.
Testing was done at 1080p using the game's "Ultra" quality preset, which maxes out all detail settings.
Let's preface the results below 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 stuttering that can 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.
To get a sense of how frame times correspond to FPS rates, check the table on the right.
We're going to start by charting frame times over one representative test run for each system. (That run is usually the middle one out of the five we ran for each card.) These plots should give us an at-a-glance impression of overall playability, warts and all. You can click the buttons below the graph to compare different cards.
From these plots alone, we can tell that the Radeon HD 7790 and GeForce GTX 650 Ti 1GB are pretty closely matched—though the GeForce does seem to exhibit more frequent latency spikes. (More on that in a moment.) We can also see that the Radeon HD 7850 2GB spends more time above the 20-ms mark than the GeForce GTX 650 Ti Boost. The contest between the 7870 and the GTX 660 is too close to call at this stage.
Now, we can slice and dice our raw frame-time data in several ways to show different facets of the performance picture. Let's start with something we're all familiar with: average frames per second. Average FPS is widely used, but it has some serious limitations. Another way to summarize performance is to consider the threshold below which 99% of frames are rendered, which offers a sense of overall frame latency, excluding fringe cases. (The lower the threshold, the more fluid the game.)
The FPS graph shows the 7790 slightly edging out the GTX 650 Ti 1GB. However, the 99th-percentile graph paints a different picture, since it accounts for those frequent spikes we spotted in the frame-by-frame plot.
There's less of a disparity between the FPS and 99th-percentile results for the other cards. The 7850 2GB clearly lags behind the GTX 650 Ti Boost, and the 7870 is on roughly even footing with the GTX 660. In all cases, the FPS results are somewhat more flattering to the Radeons than the 99th-percentile ones.
By the way, those 99th-percentile figures only capture a single point along the latency curve, but we can show you that whole curve, as well. With single-GPU configs like these, the right hand-side of the graph—and especially the last 5% or so—is where you'll want to look. That section tends to be where the best and worst solutions diverge.
The 7790's plot is interesting. 95% of the card's frame times are slightly lower than on the GTX 650 Ti 1GB, but the last 3-4% of frame times are clearly higher. The GTX 650 Ti 1GB seems to maintain consistently low latencies throughout a greater proportion of the run—hence its victory in our 99th-percentile rankings above.
The other cards don't exhibit particularly unusual behavior. The 7850 2GB is a little slower overall than the GTX 650 Ti Boost; its plot sits higher on the graph and starts to spike a little earlier. The 7870 and GTX 660 both spike at roughly the same percentage, but the 7870 maintains slightly lower frame times throughout most of the run. That explains the 7870's top spot in the FPS rankings and its parity with the GTX 660 in our 99th-percentile data.
Finally, we can rank the cards based on how long they spent working on frames that took longer than a certain number of milliseconds to render. Simply put, this metric is a measure of "badness." It tells us about the scope of delays in frame delivery during the test scenario. You can click the buttons below the graph to switch between different millisecond thresholds.
Compared to their AMD counterparts, the GeForces spend a little more time working on frames that take 50 ms or more to render. That said, the total time involved is relatively small. If we lower the threshold to 33.3 ms (which corresponds to 30 FPS), the GTX 650 Ti 1GB and GTX 650 Ti Boost perform quite a bit better than their AMD rivals, but the GTX 660 loses to the 7870. The picture at 16.7-ms (equivalent to 60 FPS)is similar, except the two slowest cards are evenly matched, no doubt because they struggle at these detail settings.
Overall, I'd say the GTX 650 Ti Boost and the 7870 win out over their respective rivals. The Boost is clearly quicker than the 7850 2GB, and its 6-ms disadvantage in the beyond-50-ms chart isn't enough to tip the balance. The 7870, meanwhile, manages equivalent 99th-percentile frame times with fewer unusual spikes than the GTX 660, so it's clearly a better choice.
As for the 7790 and GTX 650 Ti 1GB, we should test those at lower detail settings before choosing a winner.