Compare Game Old vs New Engine Performance Fairly

Comparing a game’s old engine to its new engine is about measuring the real impact of the engine change, not just the settings or hardware. This guide gives you a repeatable method to benchmark both versions under identical conditions, so you can see the actual performance difference and decide if the upgrade is worth it.

What You’ll Need

• Two installs of the game: one with the old engine, one with the new engine (if the game offers a beta branch or a separate version).
• A benchmarking tool like FRAPS, MSI Afterburner, or the built-in benchmark if the game has one.
• A stopwatch or a video recording tool (OBS, ShadowPlay) to capture timed runs.
• A notepad or spreadsheet to log your results.
• A stable, identical hardware setup (same PC, same background processes closed).

Step-by-Step

1. Set a fixed benchmark scene. Choose a specific in-game location (e.g., a busy city street, a dense forest) that is easy to reproduce. If the game has a built-in benchmark, use that. If not, pick a route you can walk or fly through in a set time, and record it so you can repeat it exactly.
2. Lock the graphics settings. Write down every graphics option (resolution, texture quality, shadow quality, anti-aliasing, etc.) and set them identically in both engines. Take screenshots of the settings menu to prove it.
3. Use the same resolution and refresh rate. Set both to the same native resolution and the same refresh rate (e.g., 2560×1440 @ 60 Hz). Do not use dynamic resolution or upscaling unless you are specifically testing that feature.
4. Close all background apps. Close browsers, Discord, streaming software, and any overlay that is not the benchmark tool. Disable Windows Game Bar and other overlays that can affect performance.
5. Run the benchmark at least three times per engine. This accounts for variance. For each run, record the average FPS, the 1% low FPS, and the frame time (if available).
6. Log the results. In your spreadsheet, note the engine version, the settings used, the date, and the three runs. Calculate the average of the three runs for each metric.
7. Compare the averages. Subtract the old engine’s average FPS from the new engine’s average FPS. Divide by the old engine’s average and multiply by 100 to get the percentage change. Do the same for 1% lows.
8. Check for visual differences. Take screenshots at the same spot in both engines. Compare image quality, but note that any visual difference is part of the engine change.
9. Test a second scene. If possible, repeat the process in a different area (e.g., an indoor area with lots of geometry) to see if the performance delta is consistent.
10. Draw a conclusion. Based on your data, decide if the new engine is faster, slower, or equal. Remember that a 5% difference might be within margin of error.

Common Mistakes

• Not resetting settings. The new engine might default to different settings (e.g., higher shadow quality) that skew the results. Always manually set them.
• Using a different scene. Comparing a city scene in the old engine to a forest scene in the new one is not a fair test.
• Ignoring background processes. A Windows update or a background download can tank FPS in one run.
• Only measuring average FPS. Average FPS can hide stuttering. Always look at 1% lows.
• Not warming up. Some games have shader compilation stutter on the first run. Do a practice run before recording.
• Changing drivers mid-test. Update your GPU driver before starting, and do not change it between engines.

Tips

• Use a fixed camera. If the game has a photo mode or a free camera, use it to get a perfectly identical view.
• Record with a hardware encoder. Use NVENC or AMD’s encoder so the recording does not affect FPS.
• Check the engine version. Verify you are actually on the old engine and not a patch that changed it.
• Test with vsync off. Vsync can cap your FPS and hide differences. Turn it off for benchmarking.
• Use a frame time graph. A flat line is better than a spiky one, even if the average is lower.

Troubleshooting

• FPS is different between runs of the same engine. This is normal. Run more times and take the median. If the variance is huge, check for thermal throttling or background processes.
• The game crashes on one engine. Try lowering settings slightly to get a stable run. If it still crashes, note it as a data point (crash counts as a failure).
• You cannot install both engines. Some games only offer the new engine. In that case, you can compare against public benchmarks from the old engine, but be aware that their hardware may differ.
• The benchmark scene is not reproducible. Use a save file or a specific time of day to force the same conditions.
• Frame time spikes. If you see spikes, check if it is shader compilation (first run) or a background process.

Frequently Asked Questions

Do I need to reinstall the game for the old engine?

Yes, if the game does not offer a way to switch engines. You may need to back up your save files and reinstall the old version from a disc or a digital storefront that still offers it.

How many runs should I do?

At least three, but five is better. More runs reduce the impact of outliers.

What is a 1% low FPS?

It is the average FPS of the worst 1% of frames. A low 1% low means the game stutters or drops frames noticeably, even if the average is high.

Can I compare on different PCs?

No, that would be an unfair test. Use the same PC for both engines.

Should I update my GPU drivers between tests?

No. Use the same driver version for both engines. Update before starting the tests.

What if the new engine has better visuals?

That is a valid result. Note it, but remember you are comparing performance, not just visuals. You can mention that the new engine looks better but runs slower, or vice versa.

Related

• Physics That Keeps Humans Valuable as AI Changes Work
• Small Modular Reactors vs Diesel for Remote Canada
• DeepSeek AI Explained: How It Works and Which Model to Use