Document P1 and P3 integrated acceptance

docs/manual/editor/diagnostics.md

@@ -37,9 +37,12 @@ Use the **Visibility** selector to compare **Direct**, **GPU frustum**, and **GP
 
 Use the **Temporal** selector for **Off**, **TAA**, or **Upscale**. Upscale shows a
 50–99% render-scale slider; output UI remains sharp. The requested/effective
-mode, fallback reason, internal extent, history reset reason and temporal GPU
-pass times are shown separately from visibility and HZB history. This selector
-only changes the live Editor viewport. See [Temporal rendering](temporal.md).
+mode, fallback reason, internal extent, clip/internal-pixel jitter, history
+reset reason and temporal GPU pass times are shown separately from visibility
+and HZB history. **Count temporal pixels (GPU readback)** optionally shows
+accepted and rejected history samples for output-scene pixels; it adds GPU
+work and is off by default. This selector only changes the live Editor
+viewport. See [Temporal rendering](temporal.md).
 
 The panel reports the previous completed frame: renderer wall time, GPU timestamp time where available, synchronous readback time, draw calls, packed vertices, culled meshes, textures, explicit Vulkan allocation sizes, actual validation availability/errors, and GPU pass-label count. It also shows whether GPU visibility ran, submitted indirect bins, visible instances, frustum rejects, deferred and post-pass visible instances, HZB history validity, counts per prepared LOD level, and GPU pass timings where available. GPU counts are explicitly marked unavailable until the first frame rendered with diagnostics open; only a displayed zero is a measured zero. **Previous HZB history: invalid** is expected after a camera cut or resize until compatible depth history is available. A current HZB preview can still exist after that first frame because it was built from the current depth. Renderer wall time includes waiting for GPU work; it is not thread CPU usage. Memory excludes driver-internal allocations. The overlay itself adds drawing work, so hide it for a baseline performance measurement.
 

docs/manual/editor/export.md

@@ -21,6 +21,11 @@ Play captures the authoring document when requested. Editing the document while
 build runs does not silently change that snapshot. Runtime movement, spawned objects,
 and gameplay progress do not write back into authoring or its Undo history.
 
+Build and Export capture gameplay sources when their worker begins. If you edit C++
+while a job waits in the queue, it builds the newer snapshot and the Inspector's
+schema status uses that snapshot's hash. An edit during the active build causes
+the job to reject changed sources; run Build again after the edit.
+
 The Player supports pause and single-step. Editor controls use a private session
 control file; they do not expose runtime entity queries through MCP. Closing the
 Player is observed by the Editor, which keeps its logs and authoring state.

docs/manual/editor/profiling.md

@@ -168,7 +168,11 @@ otherwise they are `null`. `gpu_allocated_bytes` includes live temporal targets
 and histories, subject to the allocation limits described above. Compare full
 frame GPU and renderer wall time too: scene raster savings can be offset by
 resolve, memory and synchronous readback. A valid frame-level history flag says
-the previous frame may be sampled, not that every pixel accepted it. For image
+the previous frame may be sampled, not that every pixel accepted it. Normal
+Player profiles set `temporal_counters_valid: false` and leave
+`temporal_accepted_pixels`/`temporal_rejected_pixels` null. The Editor's optional
+**Count temporal pixels (GPU readback)** toggle measures them for diagnostic
+frames but adds GPU work, so keep it off for performance runs. For image
 quality, inspect a still thin edge, a slow pan and a newly uncovered surface, and
 compare the same frame against Off. See [Temporal rendering](temporal.md) for
 mode controls and native C++ configuration.

docs/manual/editor/temporal.md

@@ -17,6 +17,15 @@ view, resize, scale switch or compatible shader reload is expected. A valid hist
 does not imply every pixel reused it: newly visible surfaces can still reject
 their individual history samples.
 
+The panel also shows the jitter in clip coordinates and the equivalent internal
+pixel offset. Turn on **Count temporal pixels (GPU readback)** only when you need
+to inspect how many output-scene pixels actually blended previous color versus
+rejected it. The counts are unavailable until a frame completes with the toggle
+on; a displayed zero is a measured zero. This optional GPU buffer clear, atomic
+count and readback add diagnostic work. Leave it off for timing comparisons.
+The toggle affects only the Editor viewport. Counting pauses while the panel is
+closed and resumes when it reopens with the checkbox still selected.
+
 For a Player or exported game, select the mode at launch:
 
 ```sh
@@ -62,4 +71,7 @@ faster. Compare Off, TAA and Upscale at the target output resolution with both
 still and moving cameras, and inspect thin objects and opening doors before
 choosing a mode for a game. The [P3 temporal validation record](https://github.com/emil28092005/Faset_Engine/blob/main/docs/validation/p3-temporal-2026-09-24/README.md)
 contains source captures, paired image measurements and a bounded 720p cost
-profile.
+profile. The [integrated three-path matrix](https://github.com/emil28092005/Faset_Engine/blob/main/docs/validation/p3-temporal-2026-09-24/matrix/integrated-ed523c6/README.md)
+also compares Direct, GPU frustum and GPU occlusion, eight open-door frames,
+and a 2× spatial reference. In that fixed thin-wire scene the 2× reference
+remains closer to the target than TAA or Upscale.