[Jaz0000]

Documents the M1 Max opening GPU hang investigation against adc0e19, including two successful diagnostic skip runs for compute shader prefix 93b8cf2a4840b5f2 and the user's later gameplay report from a combined trial build.

This is a documentation-only draft, not a root-cause rendering fix. The existing TOTK_SKIP_CS diagnostic is sufficient to reproduce the workaround; no default dispatch skipping is introduced. The document records unsuccessful controls, sampled buffer bounds, causal limitations and remaining validation. Opening-video artifact status in the latest combined trial remains unconfirmed.

Related: #2080 (separate NDVE movie experiment), #2036, #2040. No game resources, generated shaders, keys, saves or raw logs included.

Validation: locally built trial runtime and signed App passed packaging self-check; two isolated skip runs recorded no GPU Hang during their observation windows. Long-run gameplay and correctness remain unverified.

[ly05010419]

Thanks for writing this up so carefully. Since it records an investigation and a diagnostic workaround rather than a fix (as you say yourself), we'd prefer to keep it out of the repository's docs. Could you post it in #2036, or a new issue, instead? That keeps the data, the controls you ruled out and the TOTK_SKIP_CS steps in one place, where the work on the root cause can continue.