perf(generation): field-exact LOD0 overhead removal; the unknown bucket is Core.dll

Field-preserving floor round (Codex; reviewed by me against the code):
- exact per-(sample, passage) caching of native/generic floor projections and walkable-air
  checks, behind voxel.FloorRound1PassageProjectionCache
- a zero-work disabled path for VoxelDensityProfile::FScopedTimer, unchanged when enabled
- pointer / range-for iteration in the proven-hot tunnel, passage, spatial and candidate loops
  (TArray RangeCheck was ~6.5% of LOD0 samples on the Development binaries the owner plays on)
- FVector2D temporaries replaced with scalar doubles, same operations in the same order
- FSurfaceColumn made trivial, so thread-local construction does no work. Verified safe:
  ComputeSurfaceColumn writes all five outputs unconditionally before any branch, and every read
  of Box.Cols is gated by Box.Computed

Verified: export OBJ byte-identical (b3e5f4c3..9b377); capability 20,830 / 10,909; all 8
interleaved game runs gave identical totals (841 tiles / 669,834 triangles static, 2,348 /
1,985,216 moving, the same as previous rounds). Per-tile identity for the always-on changes still
needs a check against a stored trace; that is queued first in the next round.

Game path (sampler off, two interleaved runs each):
  projection-cache switch alone: neutral (static worker +0.41%, moving -0.05%)
  vs the previous round (cfad3d1, same harness; cross-round numbers reproduced within ~1% here):
    LOD0 generation p50  static 76.6 -> 65.8 ms (-14%), moving 72.4 -> 59.5 ms (-18%)
    LOD0 request p50/p95 moving 79.4/158.5 -> 67.6/137.6 ms
    worker seconds       static 39.2 -> 33.4, moving 113.7 -> 97.9
  The always-on changes carry the gain; they have no switch, so this is a cross-round comparison.

The sampler's ~12.9% "unknown" LOD0 leaf is engine code: UnrealEditor-Core.dll + RVA 0xC080,
without a Core PDB. The old FSurfaceColumn label is gone. Next round names it by its callers.

PERF-SAMPLED.md: corrected a sentence that repeated my false "daytime use" explanation.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TaFdXhTdYMkW1xrV6ApT6Y
This commit is contained in:
Fr0zkaandClaude Opus 5 committed 2026-09-13 21:00:34 +02:00
1 parent b839edd3a3
commit ab4bbe3680
9 files changed
+629 -170

No files matched your search

+15 -9
View File
@@ -206,8 +206,12 @@ private:
// uphill). See SurfaceDensityFromColumn for the warped-terrain union that makes the shelf.
struct FSurfaceColumn
{
float TerrainZ = 0.0f; float CeilSurf = 0.0f;
float OverhangAmp = 0.0f; float DirX = 0.0f; float DirY = 0.0f;
// Toutes les sorties sont écrites par ComputeSurfaceColumn avant Computed=true ; les valeurs
// par défaut ne sont donc jamais lues et leur constructeur implicite coûte dans le TLS.
// Every output is written by ComputeSurfaceColumn before Computed=true; the defaults are never
// read, and removing them keeps FSurfaceColumn trivial so TLS construction has no leaf work.
float TerrainZ, CeilSurf;
float OverhangAmp, DirX, DirY;
};
struct FSurfaceColumnBox
@@ -2355,7 +2359,6 @@ float UVoxelGenerator::GetDensityAt(float WorldX, float WorldY, float WorldZ) co
Result = GetMazeDensity(WorldX, WorldY, WorldZ, CP_Maze); break;
case ECaveGeneratorType::SurfaceWorld:
{
FSurfaceColumn Col;
// Integer XY (the density grid) → reuse the column down its whole Z extent (T1.a).
// Fractional XY (gradient-normal samples) → compute directly (no cache key).
if (WorldX == FMath::FloorToFloat(WorldX) && WorldY == FMath::FloorToFloat(WorldY))
@@ -2373,17 +2376,20 @@ float UVoxelGenerator::GetDensityAt(float WorldX, float WorldY, float WorldZ) co
Box.Cols[CI].OverhangAmp, Box.Cols[CI].DirX, Box.Cols[CI].DirY);
Box.Computed[CI] = true;
}
Col = Box.Cols[CI];
const FSurfaceColumn& Col = Box.Cols[CI];
Result = SurfaceDensityFromColumn(WorldX, WorldY, WorldZ,
Col.TerrainZ, Col.CeilSurf, Col.OverhangAmp,
Col.DirX, Col.DirY, CP_Surface);
}
else
{
float TerrainZ, CeilSurf, OverhangAmp, DirX, DirY;
ComputeSurfaceColumn(WorldX, WorldY, ChunkCoord.Z, CP_Surface, CP_BiomeCtx,
CP_SurfaceBiomeParams, CP_BiomeCache, Col.TerrainZ, Col.CeilSurf,
Col.OverhangAmp, Col.DirX, Col.DirY);
CP_SurfaceBiomeParams, CP_BiomeCache, TerrainZ, CeilSurf,
OverhangAmp, DirX, DirY);
Result = SurfaceDensityFromColumn(WorldX, WorldY, WorldZ,
TerrainZ, CeilSurf, OverhangAmp, DirX, DirY, CP_Surface);
}
Result = SurfaceDensityFromColumn(WorldX, WorldY, WorldZ,
Col.TerrainZ, Col.CeilSurf, Col.OverhangAmp, Col.DirX, Col.DirY, CP_Surface);
break;
}
case ECaveGeneratorType::VerticalShafts: