How FPS (Frames Per Second) Test / Display Benchmark Works

The benchmark uses the browser’s high-resolution timing APIs and the animation scheduling loop to sample how often frames are presented and how long each frame takes to render. Instead of focusing only on “average FPS”, it also examines frame-time spikes (jank), long frames that feel like stutter, and percentile-based “low FPS” metrics that better match what your eyes notice.

Because different displays run at different refresh rates, the tool estimates your cadence (for example near 60Hz, 120Hz, or 144Hz) and evaluates frame times relative to that target. This makes the interpretation consistent: a 16.7ms frame is “on time” for 60Hz, but it’s slow on a 120Hz panel where the target is closer to 8.3ms.

Step-by-step benchmark flow

• 1) Choose a duration: Pick how long to measure (for example 10–30 seconds). Longer tests reduce randomness and capture more real spikes.
• 2) Set warm-up time: A short warm-up stabilizes clocks, GPU boost states, and browser scheduling so the measurement is less “cold start” biased.
• 3) Select a scene load: Run a light scene to check baseline smoothness or a stress scene to expose throttling, thermal limits, and frame pacing issues.
• 4) Run the benchmark: The test records frame-to-frame timing, estimates refresh cadence, and counts long frames and likely missed refresh opportunities.
• 5) Review key metrics: See average FPS, 1%/0.1% lows, percentile frame times, jank rate, and a histogram of frame-time distribution.
• 6) Export results: Copy a plain-text report or download it so you can compare changes over time and across devices.

Key Features

Real-time FPS plus frame-time stability

Average FPS is useful, but smoothness depends on how consistent each frame is. The benchmark calculates average frame time and high-percentile frame times (like p95 and p99). It also converts those percentiles into “low FPS” values that reveal stutter patterns hidden by averages.

If your average FPS looks healthy but p99 frame time is high, you are likely experiencing intermittent hitches. That’s common with heavy tabs, background tasks, video capture, overlays, or intermittent garbage collection in web apps.

Estimated refresh rate and target frame time

Modern displays can run at 60Hz, 75Hz, 90Hz, 120Hz, 144Hz, 165Hz, 240Hz, or variable refresh modes. The tool estimates your refresh cadence from the observed frame interval distribution and uses that to set a target frame time for “on time” rendering.

This helps you interpret results correctly. For example, “90 FPS average” can feel different depending on whether your display is 60Hz (where you’re above the refresh) or 144Hz (where you’re far below the panel’s potential).

Jank and dropped-frame indicators

Not all slow frames feel the same. The tool counts frames that exceed a “jank” threshold (noticeably late compared to your typical cadence) and frames that exceed a “dropped” threshold (late enough to likely miss one or more refresh cycles). This helps you tell minor jitter from obvious stutter.

When diagnosing issues, look at jank rate together with percentiles. A small number of severe spikes can be more disruptive than many tiny deviations, and the tool makes those patterns visible.

Histogram view for quick comparisons

A frame-time histogram shows whether your performance is tightly clustered (smooth pacing) or spread out (inconsistent pacing). It’s also a fast way to compare changes: browser updates, driver updates, power settings, different cables, or a new monitor.

For a stable setup, most frames should cluster around the target interval with a short tail. A long tail suggests scheduling contention, throttling, or an overloaded render/composite pipeline.

Copyable, downloadable benchmark report

After the run, the tool generates a plain-text report with the key numbers and a short interpretation guide. You can paste it into a support ticket, keep it for a before/after record, or share it with teammates.

Because the report is plain text, it’s friendly for chat apps, issue trackers, and documentation. It’s also easy to anonymize: you can choose whether to include environment details like screen resolution and browser user agent.

Use Cases

• Verify high-refresh operation: Confirm whether your browser and display path are presenting at the expected cadence (for example 120Hz instead of falling back to 60Hz).
• Troubleshoot stutter in dashboards: Identify frame-time spikes caused by heavy charts, large DOM updates, or frequent reflows.
• Compare browsers and settings: Test the same scene in different browsers, with and without hardware acceleration, to see which configuration is smoother.
• Check laptop performance modes: Measure how battery saver, balanced, and performance profiles affect both average FPS and 1% lows.
• Evaluate external displays and docks: Detect frame pacing issues introduced by a dock, adapter, cable, or display scaling configuration.
• Assess remote desktop scenarios: Run the benchmark on a remote session to see if frame pacing is affected by the transport and encoding pipeline.
• Create before/after baselines: Document improvements after driver updates, OS upgrades, disabling overlays, or tuning browser extensions.

This benchmark is intentionally simple and repeatable. It won’t replace a full profiler, but it provides a strong first signal about whether the experience is limited by refresh rate, rendering workload, or inconsistent frame pacing.

For best results, treat it like a lab test: keep the same duration, window size, and scene load when comparing two configurations, and focus on percentiles and jank rate to judge perceived improvement.

Optimization Tips

Run tests consistently for meaningful comparisons

Close heavy background tabs, keep the browser window at the same size, and repeat the test at least twice. If your numbers vary widely between runs, extend the duration to reduce randomness and watch the percentile frame times rather than only the average. Consistency matters more than chasing a single “best run.”

Understand average FPS vs 1% lows

Average FPS can stay high even when occasional long frames ruin the experience. Use 1% low FPS and 0.1% low FPS as stutter detectors: they summarize the slowest frames you hit during the run. If those values are much lower than the average, your setup is inconsistent even if the average looks good.

Reduce sources of frame-time spikes

If you see frequent jank, try disabling heavy extensions, turning off screen recording overlays, and ensuring hardware acceleration is enabled. On laptops, plug in power and use a performance profile. If you’re using an external display, test with a different cable or port to rule out bandwidth and scaling issues.

FAQ