Profile A: official stability baseline for affected AMD PCs
This profile follows the launch workaround rather than pretending to be a universal FPS preset.
- Update AMD Software: Adrenalin Edition and reboot.
- Restore Video settings to defaults if this PC joined the Crossplay Playtest.
- Set Graphics Quality to High, as specified in the official launch workaround.
- Disable AMD Overlay and other capture overlays.
- Test the full -dx12 -USEALLAVAILABLECORES -malloc=system line; replace it with the full -dx11 -USEALLAVAILABLECORES -malloc=system line only if DX12 still crashes.
Profile B: controlled performance test
Use the same map, route and camera direction for each pass. Record display resolution, graphics preset, upscaler or reconstruction mode, VSync state, frame cap, average FPS and the lowest repeatable frame-rate event. Keep all but one variable fixed.
Tune in a safe order
First lower the overall preset by one step and compare frame time. Then test expensive effect groups before reducing texture clarity. Only change resolution or reconstruction mode after you have a stable same-resolution baseline, because those changes make before/after comparisons harder.
Protect visibility and communication
Do not optimize solely for a higher counter. Reject a profile that makes uniforms, terrain edges or map markers harder to read. Keep voice-chat volume clearly above nonessential effects; squad and leadership information has more tactical value than louder ambient sound.
What this page does not claim
No hardware-specific benchmark has been completed by this site yet, so it does not publish an invented RTX, Radeon or console preset. Future measured profiles must name CPU, GPU, driver, resolution, game build, map, capture method and sample duration.
Frequently asked questions
What settings should an AMD user try first?
Use the official stability sequence: current Adrenalin, Video defaults when relevant, High graphics, AMD Overlay off, and one complete official launch-option line at a time.
Why are there no universal Low/Medium/High values?
A defensible preset requires measured hardware and build data. This page separates verified stability steps from a controlled tuning method instead of inventing benchmark results.
Sources and verification
Sources are linked for independent checking. Dynamic figures are snapshots, not permanent values.