FCAT VR: GPU And CPU Performance in Virtual Reality

FCAT VR: An Illustrative and Accessible Tool For VR Testing

There’s a lot to like about FCAT VR, and a couple of improvements we'd like to see.

FCAT VR clearly gives us insight into graphics, host processing, and whole platform performance that we didn’t have previously, for starters. Yes, all of this data was accessible on the Vive through Valve’s SteamVR Frame Timing log and on the Rift via Event Tracing for Windows. But organizing it was a mess. We have pages of unpublished content in our CMS to show how quickly data becomes obsolete if you’re caught charting when a conclusion-changing driver is published. FCAT VR improves the rate at which we can collect data, get that information into a digestible format, and analyze it in a meaningful way. This alone should excite gamers who’ve bemoaned the lack of performance-oriented VR benchmarking.

We also have to applaud Nvidia for keeping FCAT VR freely available, modifiable, and redistributable. The very source of these tools means they’re under increased scrutiny (and rightly so). Putting them in front of the largest audience possible gives everyone an opportunity to validate their results.

The Tom’s Hardware team does enjoy one leg up. In addition to the FCAT VR tools, we also have the hardware and software infrastructure needed to intercept and record the signal coming from a graphics card before it’s displayed in a head-mounted display. Processing that information tells us the rate at which new frames are hitting the HMD. And when we overlay the FCAT VR results on top, it’s pretty easy to see how the two benchmarking techniques correlate.

There’s another interesting benefit tied to our hardware vs. software comparison: it’s possible to illustrate that 11ms isn’t an absolute threshold for rendering new frames at 90 Hz. Preemption and parallelization in the pipeline can buy precious milliseconds, and our video capture reports uninterrupted smoothness even when the software trace exposes brief incursions beyond 11ms.

In addition to its utility as a yardstick for evaluating graphics cards, FCAT VR also opens the door to evaluating other components under the duress of virtual reality. Dropping in a high-end graphics card and swapping CPUs in and out showed us that Arizona Sunshine really does respond to host processing performance—we’d like to see how other games scale based on core count and frequency. AMD’s Ryzen may give us the perfect opportunity to go into depth on the subject.

Of course, we also hope to see Nvidia continue FCAT’s development. The VR Analyzer needs more configurability, and our wish list already includes new background colors, specific dimensions for exporting charts, and an easier mechanism for altering the x- and y-axis. A reliance on runtime compatibility is worrying, but frankly somewhat unavoidable. So long as Valve and Oculus don’t deliberately impede FCAT’s functionality, future support shouldn’t be a problem. But we’ll cross that bridge when we get there. In the meantime, let us know what you think about FCAT VR, throw out some ideas for hardware comparisons, and suggest some games you’d like to see tested. We’re ready to get the data flowing.

MORE: Best Deals

MORE: Best PC Builds