AnyPS5 Progress and Game Test Milestones
A reviewed snapshot of the upstream project and selected game reports. Documentation coverage is not game compatibility.
Upstream implementation coverage
| Measure | Reported snapshot | Meaning |
|---|---|---|
| Library functions | 3,292 / 3,706 · 88.83% | Implementation coverage, not a playable-games rate. |
| Shader instructions | 1,166 / 1,166 · 100% | Recorded instruction coverage, not full rendering correctness. |
Selected game milestones
Furi · Windows source report adds menu and tutorial-fight screenshots, exact build and a repeatable post-boss crash.
Original evidence ↗Stray · New main result adds startup workarounds and a 60-second DeadCity measurement.
Original evidence ↗Ghost of Yōtei · Separate Linux/RDNA4 run records an OOM stop.
Original evidence ↗Grand Theft Auto V · A separate POC branch reports the prologue around 21 FPS on another Windows machine.
Original evidence ↗Gran Turismo 7 · Separate RTX 4090 stock-main failure report posted.
Original evidence ↗Demon’s Souls · Linux/RADV follow-up reaches character creation with visible artifacts.
Original evidence ↗Animal Well · Title, New Game and the first room render at about 58 FPS over 120 seconds.
Original evidence ↗Hades II · Native startup exits before presenting on the reported RADV validation build.
Original evidence ↗Astro Bot · Linux/RDNA4 Crash Site measurement and a separate unresolved resource-staging report are added.
Original evidence ↗Ghost of Yōtei · Windows experimental runs reach the opening scene and setup screens.
Original evidence ↗Demon’s Souls · Source update distinguishes stock-main loading failure from the patched tutorial result.
Original evidence ↗DOOM Shareware (homebrew) · E1M1 renders at 35 FPS; movement, firing, a health pickup and automap are checked.
Original evidence ↗
A merged PR or a higher implementation percentage does not establish a game’s status. Read the exact test report.
Latest AnyPS5 source and game-report updates
These changes were checked against upstream documents and original game discussions. A new report is not a completed playthrough; current main instructions are kept separate from tagged-release examples.
Lighter converter builds
Current source can build the relinker without the full runtime dependencies; game libraries still need their own build.
Apple Silicon and Rosetta
Native converter builds, x86-64 output and per-game runtime evidence are separate questions.
Furi: new Windows report and captures
Tutorial gameplay is reported on a named main build, followed by a repeatable crash after the tutorial boss.
Stray: revised Linux setup
The October 17 report separates startup workarounds, main performance and a patched test stack.
GTA V: another Windows test branch
A newer prologue report is kept separate from the older testing-branch results.
Astro Bot: Linux scene and staging notes
Add a measured RDNA4 hub scene and a separate unresolved resource-path report.
What the AnyPS5 progress snapshot establishes
AnyPS5 progress combines dated engineering work with separate per-game observations. A higher implementation counter does not mean the same percentage of games is playable.
Read AnyPS5 progress beside the original source date and build. A merged change may fix one path while another scene still needs an unimplemented function. Keep release notes separate from experimental results.
For a practical next step, choose a game report rather than treating AnyPS5 progress as a system-requirements calculator. Preserve the platform, patches and limitations that belong to that observation.