Skip to main content
  1. Home
  2. Computing
  3. Gaming
  4. Features

AMD just released an awesome new benchmark for DirectX 12 games – here’s how to use it

Add as a preferred source on Google

Recording frame times is a crucial part of benchmarking a gaming PC, but it’s become more difficult as of late because DirectX 12 doesn’t work quite the same as DX11. Older tools, most notably FRAPS, can’t record DX12 games. And while a few benchmarks exist, they’re either specific to a particular game, or difficult to use.

Luckily, AMD has decided to save the day. It commissioned a third-party developer to build an easy to use tool for DirectX 12 and Vulkan benchmarking. The result is OCAT. Though developed by AMD, it works with Intel and Nvidia graphics, as well.

Recommended Videos

We’ve started to use OCAT for benchmarking over the last week, and it looks like it’ll become our primary tool for games that lack an in-game benchmark. It’s even better than FRAPS, as it offers better formatting of data and a more detailed in-game overlay.

And getting started is simple.

The basics

You can download OCAT from Github.

Once you’ve installed OCAT, you’ll find yourself viewing a screen with a variety of squares on it. You may not even have to change these default settings, but there are a few things you should know, because OCAT’s behavior may not be what you’re used to.

ocat1
Image used with permission by copyright holder

The main point is that OCAT does not capture only a specific application by default. Instead, “Capture All Applications” is highlighted. This means what it says. All applications with a visible window will be benchmarked, from games to your web browser.

There’s nothing wrong with that if you’re doing a basic benchmark. Results are separated by application, so you shouldn’t see confusion in your numbers. However, it can clutter up the files you see, so you can instead select “Capture Single Application.” You can then select the target executable through Windows Explorer, or you can manually enter the executable location in the field to the left. You can also manually enter commandline arguments.

Check out the checkboxes

You’ll also see three checkboxes, each marked.

ocat2Write performance summary enables a spreadsheet called “perf_summary.csv” which is output to your /Documents/OCAT/Recordings folder in Windows. The summary lists average framerate, average frame time, and 99th-percentile frame time for all monitored processes. For most people, that’s all the information you’ll need.

Write detailed performance report will output a specific spreadsheet for each process monitored by OCAT. In it, you’ll find precise information about each frame generated during the test run, so those so inclined can run a custom analysis the numbers. It could be useful, for example, if you want to know how many frames took more than 33.3 seconds to generate. These reports are also found in the /Documents/OCAT/Recordings folder.

Record performance for all processes dictates if performance is recorded for all processes, or just for the active process. It’s somewhat redundant with selecting an application for capture. We found it a bit unreliable, so if you just want to capture results from a game, we recommend you use the “Capture Single Application” option instead to select the game’s executable.

Don’t forget to hit start!

This may sound silly, but it’s a bit different from how other tools work.

You need to hit start.

Otherwise, no overlay will appear, and nothing will happen when you hit the recording hotkey. This is different from FRAPS, which always shows its overlay if the application is loaded, whether you want it or not.

Also, if you’d like to record for an unlimited period of time, you’ll need to set the Recording Time to “0” seconds. By default, it’s set at 60 seconds.

A great way to benchmark DirectX 12

We tried out OCAT with Battlefield 1, Civilization IV, Deus Ex: Mankind Divided, and Warhammer: Total War, to see how well it worked.

The verdict? It’s fantastic. The reports are the easiest to read of any tool we’ve used, and the overlay is the best, too. We think this will be the tool of choice for amateur benchmarkers going forward, as well as for professionals (such as ourselves) who desire a wealth of data.

Matthew S. Smith
Matthew S. Smith is the former Lead Editor, Reviews at Digital Trends. He previously guided the Products Team, which dives…
What’s the Right Laptop for a Student on a $1,000 Budget?
The best college laptop isn't the one with the longest spec sheet. It's the one that fits your budget and the way you'll actually use it.
Dell XPS 14 Review: Sticker

This post is brought to you in paid partnership with Dell

Buying a laptop for college isn't a decision made in isolation. Tuition, textbooks, housing, meal plans, transportation, and course materials all compete for the same budget, leaving most families with one question before the semester begins: how much is enough to spend on a laptop without paying for features that may never get used?

Read more
Siri is about to hear everything you say, but Apple’s privacy approach has me cautiously on board
Apple's new Audio Intelligence features want to hear almost everything you say. A privacy document released alongside them explains why I am mostly okay with that.
Apple Watch Audio Intelligence features

Apple used its September 2026 launch event to enter a feature category it has mostly avoided until now: ambient listening. Siri Recap, Live Rewind, Music Recognition with Shazam, and Sound Recognition all landed on the Apple Watch Series 12 and Apple Watch Ultra 4. All four depend on the watch microphone picking up sound around you far more often than any previous Apple product has.

Siri Recap listens for conversations throughout your day and turns them into short AI summaries you can check later. Live Rewind is smaller in scope but arguably more useful day-to-day. It transcribes the last 15 seconds of whatever was just said with a double-press of the crown. This is perfect for catching a name, a book title, or directions you missed the first time. Music Recognition uses Shazam to identify songs playing nearby without you lifting a finger, much like Now Playing on Google's Pixel devices. Sound Recognition is a useful accessibility feature that listens for sirens, doorbells, and alarms to alert users who have a hearing impairment.

Read more
Can You Turn a Mini PC Into a Local AI Agent?
Furniture, Table, Computer

This post is brought to you in paid partnership with MSI

Not every AI task needs the scale of the cloud. An employee searching company documents, a retail kiosk answering product questions, or a digital sign reacting to customer behavior all need fast responses, but they don't necessarily need to send every prompt to a remote data center. Running those workloads locally reduces latency, keeps sensitive information closer to where it's generated, and can lower the ongoing cost of AI deployments. As a result, many organizations are moving toward hybrid AI architectures that handle routine requests on-device while reserving cloud models for tasks that genuinely need more processing power.

Read more