[oliverames]

What and why

Records what iPhone Duo runs (iPhone19,4) from 2026-11-07 to 11-10 showed, in the two game notes, so others can pick up the open problems without repeating the experiments. Notes only: no code, no switch changes.

• Mario Kart World, update 1.8.0: the slow-motion stretch after about 30 to 70 s, the GPU clock step the canary (DM_GPU_CANARY=1 (opt-in): a fixed compute dispatch once a second, to tell a GPU clock fall from a heavier scene #1709) reads, what did not change it (HDR, present interval, the two pacing switches of Opt-in pacing switches for the iPhone judder: DM_PCORE, DM_RT60, DM_ITV_LATCH (not fully tested) #1713, since closed, a cold start), why DM_PRESENT_HZ=30 is not a workaround, that DM_FPS_TARGET was not applied on 1.8.0 on those builds (main's 6815810df of Nov 12 places the frame step on 1.8.0, not yet run on an iPhone; the note says so), and the 2026-11-09 runs on a build from main 04811a63.
• Ocarina of Time: replaces the stale "no iOS shader build or iPhone run" line with the iPhone result: no memory wall, the foliage that rendered far too bright on 2026-11-07 renders right on a build from main 04811a63 (2026-11-09), and the frame rate in play is low and open.

Not fully tested: the notes report device runs that were read from logs; the causes are not found.

Checked

• python3 dm/tools/tests/test_docs.py passes
• python3 dm/tools/docs_symptoms.py runs clean
• scripts/regress.py check --games not run (docs only)

Front ends

No front-end interface change.

Docs

dm/games/0400C3F00006E000/README.md (Open) and dm/games/0400486024DEC000/README.md (Open).