engine(macos-arm64): publish a967d12b - parallel draw-list replay (Vulkan) + parallel draw-list recording (all drivers); GodotSharp regenerated

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_018hSLRwirUk2NF4j1v2kH4Z

README.md

@@ -31,10 +31,10 @@ into GodotSharp, so GodotSharp had to be regenerated — and must ship here alon
 
 | Platform | Engine | Terrain3D | GodotSharp shipped? |
 |---|---|---|---|
-| `engine/macos-arm64/` | **4.7.2** (`e64890fd`) — `4fde0e94` plus the Apple render-thread autorelease drain (fixes the Mach-port / Metal leak under `thread_model=2`) and the G4 mergeable-draw counter | built-in module | ✅ yes (`GodotSharp/`) |
+| `engine/macos-arm64/` | **4.7.2** (`a967d12b`) — `e64890fd` plus parallel draw-list replay (Vulkan secondary command buffers) and parallel draw-list recording (Forward+ render lists over worker threads, every driver) | built-in module | ✅ yes (`GodotSharp/`) |
 | `engine/windows-x64/`  | **4.7.2** (`e64890fd3`) — same hash as macos-arm64: `4fde0e94` (G2 pooled cull, fixed shadow band, `STATIC_SHADOW_AT_REST`) plus the G4 mergeable-draw counter (the autorelease-drain fix is Apple-only); `template_release` here is still the `1e7be28b` tree | built-in module | ✅ yes (`GodotSharp/`) |
 
-Mac is engine `e64890fd` as of 2026-10-18 (`--version` prints `4.7.2.stable.mono.custom_build.e64890fd3`): `4fde0e94` plus the separate render thread draining an Objective-C autorelease pool each flush — before, autoreleased Metal command buffers, encoders and drawables leaked ~88 Mach ports/s until no new Metal process could start (66k ports / 23 GB after 1.5 h); now flat at ~640 ports — and `RENDERING_INFO_GPU_CULL_MERGEABLE_DRAWS_IN_FRAME`. Underneath, `4fde0e94` added: G2 pooled GPU cull over shared storage arenas (`rendering/occlusion_culling/gpu_cull_pooled`, game dial `env gpucull pool`), a fixed shadow LOD band for gpu-culled MultiMeshes (`gpu_cull_shadow_fixed_band`, `env gpucull shadowband`) and the `STATIC_SHADOW_AT_REST` shader built-in (`static_casters_at_rest`, `env gpucull staticrest`); all default off. GodotSharp regenerated (API + Tools). Oracles on Metal: `multimesh_cull_pool` 4524/4524, `draw_indirect_count` 24/24. Windows is also engine `e64890fd` as of 2026-10-18 (`.agents/handoffs/forge-windows-e64890fd.md`) — both platforms and the `4.7.2@e64890fd3` pin share one hash; the autorelease-drain fix is Apple-only, Windows gains the G4 census counter (`env gpucull status` shows `(N mergeable)`). Headless game boot clean on Forge. Underneath, `4fde0e94` (`.agents/handoffs/forge-windows-g2-4fde0e94.md`) added: G2 pooled GPU cull over shared storage arenas (`rendering/occlusion_culling/gpu_cull_pooled`, game dial `env gpucull pool`), a fixed shadow LOD band for gpu-culled MultiMeshes (`gpu_cull_shadow_fixed_band`, `env gpucull shadowband`) and the `STATIC_SHADOW_AT_REST` shader built-in (`static_casters_at_rest`, `env gpucull staticrest`; the game's `tree_wind.gdshader` uses it); all default off. Underneath is the `9dcc5d458` pair (SSCS contact shadows, hero shadow min-combine, upstream 4.8 shadow fixes #121775/#124348 with the game sun bias 0.014, `draw_list_draw_indirect_count`), where the D3D12 oracle on RTX 5080 was 22 passed / 2 failed — the base-instance pair, the documented D3D12 gap (`.agents/handoffs/forge-windows-sscs-9dcc5d45.md`); `--gpu-validation` needs the D3D12 debug layer, which Forge lacks (Graphics Tools not installed). Still owed on Windows: `multimesh_cull_pool` oracle (gates making pool the default), the validation pass, and in-world checks (`herocombine`, `contactshadow`, SSCS in `--gpu-profile`, sun acne at bias 0.014, pooled gpucull census, mergeable counter).
+Mac is engine `a967d12b` as of 2026-10-19 (`--version` `4.7.2.stable.mono.custom_build.a967d12bf`): `e64890fd` plus parallel draw-list replay (Vulkan only) and parallel draw-list recording (every driver), both engine-default off and turned on by the game; GodotSharp regenerated; oracle `parallel_draw_replay` Vulkan 10/10, Metal 5/5 (`engine/.agents/wiki/parallel-draw-replay.md`). Windows to follow (`.agents/handoffs/forge-windows-parallel-replay.md`). Before that, Mac was engine `e64890fd` as of 2026-10-18 (`--version` prints `4.7.2.stable.mono.custom_build.e64890fd3`): `4fde0e94` plus the separate render thread draining an Objective-C autorelease pool each flush — before, autoreleased Metal command buffers, encoders and drawables leaked ~88 Mach ports/s until no new Metal process could start (66k ports / 23 GB after 1.5 h); now flat at ~640 ports — and `RENDERING_INFO_GPU_CULL_MERGEABLE_DRAWS_IN_FRAME`. Underneath, `4fde0e94` added: G2 pooled GPU cull over shared storage arenas (`rendering/occlusion_culling/gpu_cull_pooled`, game dial `env gpucull pool`), a fixed shadow LOD band for gpu-culled MultiMeshes (`gpu_cull_shadow_fixed_band`, `env gpucull shadowband`) and the `STATIC_SHADOW_AT_REST` shader built-in (`static_casters_at_rest`, `env gpucull staticrest`); all default off. GodotSharp regenerated (API + Tools). Oracles on Metal: `multimesh_cull_pool` 4524/4524, `draw_indirect_count` 24/24. Windows is also engine `e64890fd` as of 2026-10-18 (`.agents/handoffs/forge-windows-e64890fd.md`) — both platforms and the `4.7.2@e64890fd3` pin share one hash; the autorelease-drain fix is Apple-only, Windows gains the G4 census counter (`env gpucull status` shows `(N mergeable)`). Headless game boot clean on Forge. Underneath, `4fde0e94` (`.agents/handoffs/forge-windows-g2-4fde0e94.md`) added: G2 pooled GPU cull over shared storage arenas (`rendering/occlusion_culling/gpu_cull_pooled`, game dial `env gpucull pool`), a fixed shadow LOD band for gpu-culled MultiMeshes (`gpu_cull_shadow_fixed_band`, `env gpucull shadowband`) and the `STATIC_SHADOW_AT_REST` shader built-in (`static_casters_at_rest`, `env gpucull staticrest`; the game's `tree_wind.gdshader` uses it); all default off. Underneath is the `9dcc5d458` pair (SSCS contact shadows, hero shadow min-combine, upstream 4.8 shadow fixes #121775/#124348 with the game sun bias 0.014, `draw_list_draw_indirect_count`), where the D3D12 oracle on RTX 5080 was 22 passed / 2 failed — the base-instance pair, the documented D3D12 gap (`.agents/handoffs/forge-windows-sscs-9dcc5d45.md`); `--gpu-validation` needs the D3D12 debug layer, which Forge lacks (Graphics Tools not installed). Still owed on Windows: `multimesh_cull_pool` oracle (gates making pool the default), the validation pass, and in-world checks (`herocombine`, `contactshadow`, SSCS in `--gpu-profile`, sun acne at bias 0.014, pooled gpucull census, mergeable counter).
 
 Long-term: CI (Gitea Actions) builds on push and publishes here / to S3 —
 `docs/ai/plans/build-pipeline-architecture-2026-06-07.md`.

manifest.json

@@ -4,16 +4,16 @@
     "macos-arm64": {
       "binary": "engine/macos-arm64/tinqs.macos.editor.arm64.mono",
       "godotsharp": "engine/macos-arm64/GodotSharp",
-      "version": "4.7.2.stable.mono.custom_build.e64890fd3",
+      "version": "4.7.2.stable.mono.custom_build.a967d12bf",
       "engine_repo": "tinqs/engine",
-      "engine_commit": "e64890fd",
-      "sha256": "ee3f58dd973f571599eed5f3d043c1b8c84bc7ee38b0e180ea8d31523e0b932b",
+      "engine_commit": "a967d12b",
+      "sha256": "abc6dc79990ad53949b35bee0f5a2029a0c0b49803c25ae3b3b792c0d233f563",
       "target": "editor",
-      "built_on": "2026-10-18",
-      "built_by": "Ozan/Kraken via Claude (render-thread autorelease drain on Apple + G4 mergeable-draw counter)",
+      "built_on": "2026-10-19",
+      "built_by": "Ozan/Kraken via Claude (parallel draw-list replay + recording)",
       "toolchain": "Xcode (Apple Silicon), SCons, VulkanSDK 1.4.350.0",
       "scons_flags": "platform=macos target=editor arch=arm64 module_mono_enabled=yes vulkan_sdk_path=~/VulkanSDK/1.4.350.0/macOS",
-      "note": "Mac. Engine e64890fd = 4fde0e94 + Apple render-thread fix: the separate render thread (thread_model=2, the game default) now drains an Objective-C autorelease pool each flush - before, autoreleased Metal command buffers/encoders/drawables leaked ~88 Mach ports/s until Metal could not start new processes (66k ports / 23 GB after 1.5 h); now flat at ~640. Also RENDERING_INFO_GPU_CULL_MERGEABLE_DRAWS_IN_FRAME (G4 census) and draw_indirect_count null scratch now errors loudly. GodotSharp regenerated (with Tools/). Oracles on Metal: multimesh_cull_pool 4524/4524, draw_indirect_count 24/24. Windows at 4fde0e94; Forge to rebuild e64890fd (handoff forge-windows-e64890fd.md) - the leak fix is Apple-only, Windows gains only the counter."
+      "note": "Mac. Engine a967d12b = e64890fd + parallel draw-list replay (06ecc5df: large draw lists recorded into Vulkan secondary command buffers on worker threads at frame end; Vulkan only, no effect on Metal) + parallel draw-list recording (Forward+ render lists split over worker threads into RenderingDevice sub-lists; every driver) + PipelineHashMapRD RWLock. Project settings rendering/rendering_device/parallel_draw_{replay,recording}/* (engine default off; the game turns both on). GodotSharp regenerated (with Tools/). Oracle tests/gpu_oracle/parallel_draw_replay: Vulkan/MoltenVK 10/10, Metal 5/5 pixel-identical. Windows at e64890fd3; Forge to build a967d12b (handoff forge-windows-parallel-replay.md)."
     },
     "windows-x64": {
       "binary": "engine/windows-x64/tinqs.windows.editor.x86_64.mono.exe",
@@ -29,7 +29,7 @@
       "built_by": "Ozan/Forge via ZCode (forge-windows-e64890fd handoff)",
       "toolchain": "MSVC 14.3 (VS2022 Build Tools), Windows SDK 10.0.26100, SCons 4.10.1, .NET SDK 10.0.103",
       "scons_flags": "platform=windows target=editor module_mono_enabled=yes -j32",
-      "note": "Windows carries e64890fd3 — SAME hash as macos-arm64 (built at e64890fd3b): 4fde0e94 G2 pooled cull pair (see its note in git history) PLUS e64890fd3b — Apple-only render-thread autorelease drain (no Windows effect), RENDERING_INFO_GPU_CULL_MERGEABLE_DRAWS_IN_FRAME (new enum value 13, G4 census — env gpucull status shows '(N mergeable)'), Metal draw_indirect_count null scratch errors loudly (no Windows effect). GodotSharp regenerated from this binary — Api AND Tools. Headless game boot clean (0 binding errors); in-world smoke (host world, 0 wrapper errors, gpucull mergeable counter) per handoff still owed. Owed on Windows (carried): multimesh_cull_pool oracle on D3D12, --gpu-validation pass (Graphics Tools not installed on Forge), in-world gpucull census + pooled on/off, tree_wind compile check, crash-report checklist (8 steps). Editor target only — template_release still 1e7be28b. D3D12 Agility DLLs not shipped (Vulkan game)."
+      "note": "Windows carries e64890fd3 \u2014 SAME hash as macos-arm64 (built at e64890fd3b): 4fde0e94 G2 pooled cull pair (see its note in git history) PLUS e64890fd3b \u2014 Apple-only render-thread autorelease drain (no Windows effect), RENDERING_INFO_GPU_CULL_MERGEABLE_DRAWS_IN_FRAME (new enum value 13, G4 census \u2014 env gpucull status shows '(N mergeable)'), Metal draw_indirect_count null scratch errors loudly (no Windows effect). GodotSharp regenerated from this binary \u2014 Api AND Tools. Headless game boot clean (0 binding errors); in-world smoke (host world, 0 wrapper errors, gpucull mergeable counter) per handoff still owed. Owed on Windows (carried): multimesh_cull_pool oracle on D3D12, --gpu-validation pass (Graphics Tools not installed on Forge), in-world gpucull census + pooled on/off, tree_wind compile check, crash-report checklist (8 steps). Editor target only \u2014 template_release still 1e7be28b. D3D12 Agility DLLs not shipped (Vulkan game)."
     }
   }
 }