DLSS Multi Frame Generation and optional neural rendering for Wuthering Waves.
Unlock RTXMFG 3x / 4x / 5x / 6x through GPU-name spoofing, an automatic RTXMFG startup patch, and NVIDIA profile tweaks. Choose your multiplier in WuWa's graphics settings; the toolkit handles setup, runtime verification, and restore.
- In-game MFG control. Follow game mode keeps the multiplier under WuWa's control.
- Guided setup. Finds Steam installs, checks compatibility, patches RTXMFG locally, and backs up changes. No separate Python or Profile Inspector installation.
- Optional neural rendering. Run NVIDIA's neural engine alongside MFG, with an effect toggle and adjustable processing resolution.
- Reversible changes. Separate install, diagnostics, and restore actions for frame generation and the neural add-on.
The validated setup is Windows x64 · RTX 4070 Ti · NVIDIA driver 616.92 · WuWa on Steam. Runtime readings confirmed 3x and 6x; 4x and 5x were tester-reported.
Setup also checks exact NVIDIA cache builds and refuses unvalidated combinations. Other RTX cards are not currently supported by this installer. Full requirements and test results →
Important
This is an experimental third-party mod. Anti-cheat approval and account safety are not established. Setup changes two global NVIDIA FG settings, which can affect other games. Review what changes.
- Download the latest toolkit ZIP and extract the entire folder.
- Close WuWa. Run Setup.cmd → 1: Install MFG unlock, approve the Windows prompt, and review the detected game folder.
- Choose the offered RTX 5080 name spoof if WuWa only shows FG on/off. Type
INSTALLto apply. Restart Windows if you enabled the spoof. - Launch WuWa and select a frame-generation multiplier in its graphics settings. Use Setup.cmd → 2: Verify while playing to check actual runtime counts.
Detailed installation · Setup menu reference · Troubleshooting
The optional OptiScaler NR integration adds NVIDIA's neural effect alongside MFG. It requires a local source build and a separately supplied compatible NVIDIA runtime. Guided setup starts the effect off; F8 toggles it after enabling. Resolution and transition stability are covered in the neural setup guide.
The custom student model is research only. Its first WuWa trial showed washed-out output, weak dark-scene accuracy, and reported stutters. The test setup has returned to NVIDIA. Research status and live-switch controls →
Does setup actually apply the custom patch? Yes. It downloads the pinned
RTXMFG v1.3.3 DLL, applies the binary edit, verifies its hash, and installs it as
winmm.dll. No manual source patch is needed. How it works →
Does 6x mean six times the responsiveness? No. It means up to five generated frames per rendered frame. FPS and input responsiveness are different; verify runtime counts and judge motion separately.
Is this official DLSS 5? No. This is a community toolkit. NVIDIA runtimes, model weights, and compiled OptiScaler DLLs are not bundled.
Close WuWa, run Uninstall.cmd, review the target, and type
UNINSTALL. It restores the managed neural add-on first, then MFG, saved NVIDIA
settings, and any GPU-name aliases. Backups are retained.
For the separate student integration, follow its restore procedure first. Restore details →
| I want to… | Read |
|---|---|
| Set up MFG or undo changes | Installation |
| Add the NVIDIA neural effect | Neural rendering |
| Fix a problem | Troubleshooting |
| Understand compatibility and measured results | Validation |
| Find source, tools, or experiments | Repository map · Research index |
Bug reports, reproducible compatibility reports, and focused improvements are welcome. See CONTRIBUTING.md for development setup, tests, and reporting guidelines. All documentation →
Built on RTXMFG, OptiScaler Neural Rendering, and NVIDIA NVAPI, with driver-profile research from FirstEverTech.
Toolkit/setup code and RTXMFG changes: MIT. OptiScaler-derived patches: GPL-3.0. Research files carry their own notices; see third-party attribution for details.