profile: in-process SIGPROF sampler + exclusive route/phase census (L77); closes to 0.1%

Opt-in (L77 build + --linux77-profile; compiled out by default, 0.02% handler cost).
FI/TN cost account closes against direct grid-fill timing within 0.15% (LOD0-2 bins within 1.4%).
Passage carving = 66% of LOD0-2 CPU (52-54% whole view), passage support 10-14%, base packed
+ analytic field ~3%. Hotspots: value noise 17%, lower_bound in VisitNearestPassageSpatialNode
14%, ApplyResolvedStratePassageCarves 6.5%. LOD0/1 complete LAST (4.1/4.6 s) while LOD2 is
ready at 2.2 s.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016dhxM43dnWvUu2B5Ph1iG1

Standalone/Source/VoxelForge/Private/VoxelMarchingCubesMesher.cpp

@@ -5,6 +5,7 @@
 #include "VoxelHybridSharpMesher.h"
 #include "MarchingCubesTables.h"
 #include "StandaloneShapeWorld.h"
+#include "VoxelForgeCore/Linux77Profiler.h"
 #include "VoxelDensityProfile.h"
 #if WITH_EDITOR
 #include "VoxelRedefinedDensityField.h"
@@ -1047,6 +1048,8 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
     // Bande de strate : seules les rangées Z réellement lues (cellules CzLo..CzHi + marges de
     // gradient) sont échantillonnées — le reste du tampon reste non initialisé et non lu.
     const uint64 StageDensityGridStartCycles=OutStageTimings?FPlatformTime::Cycles64():0;
+    VoxelForgeCore::Linux77Profiler::FScopedStage l77GridFillStage(
+        VoxelForgeCore::Linux77Profiler::EStage::GridFill);
     VoxelDensityProfile::FScopedTimer MesherDensityGridTimer(
         VoxelDensityProfile::EBucket::MesherDensityGrid);
 #if WITH_EDITOR
@@ -1264,6 +1267,7 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
                 const int32 SampleZ=OriginVoxels.Z+gz*Step;
                 const bool bBoundarySample=gx<0||gx>GridDim||gy<0||gy>GridDim
                     ||gz<0||gz>GridDim;
+                VoxelForgeCore::Linux77Profiler::FScopedHalo l77Halo(bBoundarySample);
                 VoxelForgeCore::floating_island_detail::FScopedIslandFieldCaller islandCaller(
                     bBoundarySample
                         ?VoxelForgeCore::floating_island_detail::EIslandFieldCaller::Boundary
@@ -1272,6 +1276,11 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
                     ?ShapeWorldContext->DensityAtLattice(SampleX,SampleY,SampleZ)
                     :Generator->GetDensityAtLattice(SampleX,SampleY,SampleZ,
                         bCapturePreEditGrid ? &PreEditDensity : nullptr);
+                if(OutStageTimings)
+                {
+                    if(bBoundarySample)++OutStageTimings->DensityGridHaloSamples;
+                    else ++OutStageTimings->DensityGridInteriorSamples;
+                }
                 if (bCapturePreEditGrid)
                 {
                     OutPreEditDensityGrid->Samples[LocalIndex] = PreEditDensity;
@@ -1304,6 +1313,7 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
             float Densities[8]{};
             const bool bBoundarySample=GX0<0||GX0+BatchCount-1>GridDim
                 ||gy<0||gy>GridDim||gz<0||gz>GridDim;
+            VoxelForgeCore::Linux77Profiler::FScopedHalo l77Halo(bBoundarySample);
             VoxelForgeCore::floating_island_detail::FScopedIslandFieldCaller islandCaller(
                 bBoundarySample
                     ?VoxelForgeCore::floating_island_detail::EIslandFieldCaller::Boundary
@@ -1314,6 +1324,14 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
             {
                 ++OutStageTimings->X8DensityCalls;
                 OutStageTimings->X8ActiveLanes+=static_cast<uint64>(BatchCount);
+                for(int32 Lane=0;Lane<BatchCount;++Lane)
+                {
+                    const int32 SampleGX=GX0+Lane;
+                    if(SampleGX<0||SampleGX>GridDim||gy<0||gy>GridDim
+                        ||gz<0||gz>GridDim)
+                        ++OutStageTimings->DensityGridHaloSamples;
+                    else ++OutStageTimings->DensityGridInteriorSamples;
+                }
             }
             for(int32 Lane=0;Lane<BatchCount;++Lane)
             {
@@ -1350,12 +1368,24 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
                 float Densities[CHUNK_SIZE+3]{};
                 const bool bBoundarySample=gx<0||gx>GridDim||gy<0||gy>GridDim
                     ||GzLo<0||GzHi>GridDim;
+                VoxelForgeCore::Linux77Profiler::FScopedHalo l77Halo(bBoundarySample);
                 VoxelForgeCore::floating_island_detail::FScopedIslandFieldCaller islandCaller(
                     bBoundarySample
                         ?VoxelForgeCore::floating_island_detail::EIslandFieldCaller::Boundary
                         :VoxelForgeCore::floating_island_detail::EIslandFieldCaller::GridFill);
                 ActiveShapeWorld->DensityAtLatticeZRun(SampleX,SampleY,
                     OriginVoxels.Z+GzLo*Step,Step,BatchCount,Densities);
+                if(OutStageTimings)
+                {
+                    for(int32 Lane=0;Lane<BatchCount;++Lane)
+                    {
+                        const int32 LocalGz=GzLo+Lane;
+                        if(gx<0||gx>GridDim||gy<0||gy>GridDim
+                            ||LocalGz<0||LocalGz>GridDim)
+                            ++OutStageTimings->DensityGridHaloSamples;
+                        else ++OutStageTimings->DensityGridInteriorSamples;
+                    }
+                }
                 for(int32 Lane=0;Lane<BatchCount;++Lane)
                 {
                     const int32 LocalGz=GzLo+Lane;
@@ -1505,6 +1535,7 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
                 const int32 SampleY=OriginVoxels.Y+gy*Step;
                 const int32 SampleZ=OriginVoxels.Z+gz*Step;
                 const bool bBoundarySample=gx<0||gx>GridDim||gy<0||gy>GridDim;
+                VoxelForgeCore::Linux77Profiler::FScopedHalo l77Halo(bBoundarySample);
                 VoxelForgeCore::floating_island_detail::FScopedIslandFieldCaller islandCaller(
                     bBoundarySample
                         ?VoxelForgeCore::floating_island_detail::EIslandFieldCaller::Boundary
@@ -1527,6 +1558,9 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
         }
     }
     MesherDensityGridTimer.End();
+    l77GridFillStage.Restore();
+    VoxelForgeCore::Linux77Profiler::FScopedStage l77MarchingCubesStage(
+        VoxelForgeCore::Linux77Profiler::EStage::MarchingCubes);
     if(OutStageTimings)
         OutStageTimings->DensityGridMilliseconds=static_cast<double>(
             FPlatformTime::Cycles64()-StageDensityGridStartCycles)
@@ -2119,6 +2153,9 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
     const int32 CapSurfaceTriangleCount=CapTris.Num()/3;
     TArray<uint8> GroundSkirtFaceMasks,CapSkirtFaceMasks;
     const uint64 SkirtStageStartCycles=OutStageTimings?FPlatformTime::Cycles64():0;
+    l77MarchingCubesStage.Restore();
+    VoxelForgeCore::Linux77Profiler::FScopedStage l77SkirtStage(
+        VoxelForgeCore::Linux77Profiler::EStage::Skirt);
     VoxelForgeCore::floating_island_detail::FScopedIslandFieldCaller skirtFieldCaller(
         VoxelForgeCore::floating_island_detail::EIslandFieldCaller::Skirt);
     if (bGenerateSkirts && (GroundTris.Num() + CapTris.Num()) > 0)
@@ -2382,6 +2419,9 @@ FVoxelMeshData UVoxelMarchingCubesMesher::GenerateMesh(FIntVector OriginVoxels,
         OutStageTimings->SkirtMilliseconds=static_cast<double>(
             FPlatformTime::Cycles64()-SkirtStageStartCycles)
                 *FPlatformTime::GetSecondsPerCycle64()*1000.0;
+    l77SkirtStage.Restore();
+    VoxelForgeCore::Linux77Profiler::FScopedStage l77MesherFinalStage(
+        VoxelForgeCore::Linux77Profiler::EStage::MarchingCubes);
 
     // F17 — concatène sol PUIS sky-cap : un run d'indices contigu par polygroup (voir
     // FVoxelMeshData::NumCeilingTriangles). Mêmes triangles qu'avant, seul l'ordre change.