DLSS 5 Swapper FPS: Measure Before You Tune

Establish the result you want

Decide whether the problem is low rendered FPS, uneven frame delivery, slow input response, excessive heat, or a visual trade-off. These are different outcomes. A higher displayed frame counter cannot answer all of them.

The publisher describes additional processing cost for Multipass. Feeder’s configuration documentation also describes working-resolution settings. Neither source establishes a universal performance increase for your game.

Make a baseline you can repeat

Choose a saved scene or a repeatable benchmark sequence. Fix resolution, graphics preset, display mode, frame cap, and background applications. Allow loading or shader preparation to settle before recording. Repeat the same measurement duration several times and retain the range, rather than selecting the best single run.

Record	Baseline	One changed configuration
GPU, driver, operating system	Your actual values	Keep fixed
Game build, renderer, resolution	Your actual values	Keep fixed
Swapper route and component versions	Original / disabled state	Exact installed versions
Preset and individual setting	Baseline value	One changed value
Scene and duration	Repeatable sequence	Same sequence
Rendered FPS and frame-time behavior	Measured result	Measured result
Generated output FPS, if applicable	Separate measurement	Separate measurement
Image artifacts, response, stability	Observations	Observations

Fill the table with your measurements. If your counter cannot distinguish rendered from generated frames, record that limitation alongside the result.

Tune the component that actually costs time

For Multipass, compare the lowest supported pass count with one higher count before adding more passes. More processing is not an FPS multiplier. For a documented Feeder DX11 configuration, work_resolution and work_upscale concern its working image; do not relabel them as the game’s standard DLSS Quality modes.

Follow the component’s current instructions before changing those parameters. Keep game resolution and other settings unchanged during that comparison. If a new setting damages image stability or response, return to the prior value even if a counter is higher.

Use the component map and Multipass guide to establish which control applies to your route.

Check monitoring and generated frames separately

Monitoring overlays can themselves interact with injection. If a problem occurs only when an overlay is running, repeat a controlled test with that overlay closed and record the difference. Do not assume it is the cause of every failure.

For frame-generation claims, inspect rapid camera turns, HUD elements, and input response. Our MFG explanation identifies the difference between a neural pass and generated frames. A video demonstration that changes several settings at once cannot isolate the contribution of one control.

Save a useful conclusion

A useful result names the complete setup and shows what improved, what worsened, and how much the runs varied. A preset from another GPU or game cannot predict the gain on your system. This guide provides a measurement method; it contains no independently measured FPS results.

When tuning a modded game such as The Sims 4, keep the game build, renderer, installed mods, and scene fixed. To investigate heat, also record GPU temperature, utilization, power, and ambient conditions. Compare the same workload before deciding whether a change has introduced a problem.

If the setup becomes unstable, preserve diagnostics and restore the game to recheck the baseline. If even the baseline fails, use crash diagnosis before tuning further.