Fr0zka/VoxelForge
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

standalone: per-tile speed at plugin level, streaming like the game, softer lighting

Browse Source 
Cause found by profiling: room/SDF graph-cache churn (exact XY bounds + sample order rebuilt it
constantly). Region grouping (one LOD3 tile: 1,084 -> 229 builds), taller-tile traversal for LOD6
(36,480 -> 12,160), extent-dependent cache slots, lambdas instead of recursive std::function.
Serial ms/tile, 64 nearest tiles per level (before -> after):
  L0 50.4 -> 50.6, L1 112 -> 110, L2 24 -> 22, L3 223 -> 73, L4 269 -> 134, L5 393 -> 127, L6 283 -> 145.
Hashes of all levels unchanged; parity 3,436/3,436 bit-identical.
Full clipmap, 16 workers BelowNormal: full detail 66.6 s, all 12,054 tiles 175 s. In-stream L0 tail is
high (median 83 ms, mean 198 ms, p95 669 ms) vs 51 ms serial.
Streaming: recentre keeps the intersection, generates only missing tiles nearest/finest first;
one crossing = exactly 441 new tiles in 6.1 s, no stale results.
Lighting: normals already come from the density gradient; the striping was the squared headlamp;
broad linear fill + warmer palette. Daytime CPU rule documented in Standalone/README.md.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MmSoF64BbRFmk9C3p1VsBJ
This commit is contained in:
Fr0zkaandClaude Opus 5.5 committed 2026-10-03 06:49:41 +02:00
1 parent 3f300b912a
commit 81b7d6d0ff
24 files changed
+1096 -196
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

Standalone/Source/StreamingWorld.h
+45 -9
View File
@@ -3,6 +3,7 @@
#include "WorldMesher.h"
#include <condition_variable>
#include <array>
#include <unordered_set>
struct FStandaloneTileKeyHash
@@ -30,6 +31,7 @@ struct FStandaloneCpuTile
std::vector<uint32> Indices;
uint64 DensitySamples=0;
double WorkerSeconds=0.0;
std::array<uint64,VoxelDensityProfile::BucketCount> ProfileCycles{};
};
struct FStandaloneStreamingUpdate
@@ -40,6 +42,7 @@ struct FStandaloneStreamingUpdate
int32 NewJobs=0;
int32 CancelledQueuedJobs=0;
int32 FullResTiles=0;
std::vector<std::pair<uint64,uint64>> CancelledByCrossing;
};
struct FStandaloneTileJob
@@ -48,6 +51,20 @@ struct FStandaloneTileJob
int32 BandZMin=MIN_int32;
int32 BandZMax=MAX_int32;
uint64 RequestedCrossing=0;
double PriorityDistanceSquared=0.0;
};
struct FStandaloneTileJobHeapOrder
{
bool operator()(const FStandaloneTileJob& A,const FStandaloneTileJob& B)const
{
if(A.PriorityDistanceSquared!=B.PriorityDistanceSquared)
return A.PriorityDistanceSquared>B.PriorityDistanceSquared;
if(A.Key.Level!=B.Key.Level)return A.Key.Level>B.Key.Level;
if(A.Key.Coord.Z!=B.Key.Coord.Z)return A.Key.Coord.Z>B.Key.Coord.Z;
if(A.Key.Coord.Y!=B.Key.Coord.Y)return A.Key.Coord.Y>B.Key.Coord.Y;
return A.Key.Coord.X>B.Key.Coord.X;
}
};
namespace StandaloneWorldMesher
@@ -99,6 +116,7 @@ public:
explicit FStandaloneStreamingGenerator(FStandaloneOwnerWorld& InWorld,int32 InWorkerCount=6)
:World(&InWorld),WorkerCount(FMath::Clamp(InWorkerCount,1,32))
{
VoxelGenLOD::InitializeConsoleSwitches();
UVoxelMarchingCubesMesher::InitializeConsoleSwitches();
}
FStandaloneStreamingGenerator(const FStandaloneStreamingGenerator&)=delete;
@@ -153,12 +171,20 @@ public:
Desired.clear();Desired.reserve(NewDesired.Tiles.size()*2+1);
for(const FVoxelTileKey& Key:NewDesired.Tiles)Desired.insert(Key);
FullResTileCount=Stats.FullResTiles;
std::unordered_map<uint64,uint64> CancelledByCrossing;
auto QueueEnd=std::remove_if(Queue.begin(),Queue.end(),[&](const FStandaloneTileJob& Job)
{
if(Desired.contains(Job.Key))return false;
Outstanding.erase(Job.Key);++Stats.CancelledQueuedJobs;return true;
Outstanding.erase(Job.Key);++Stats.CancelledQueuedJobs;
++CancelledByCrossing[Job.RequestedCrossing];return true;
});
Queue.erase(QueueEnd,Queue.end());
Stats.CancelledByCrossing.reserve(CancelledByCrossing.size());
for(const auto& Pair:CancelledByCrossing)Stats.CancelledByCrossing.push_back(Pair);
for(FStandaloneTileJob& Job:Queue)
Job.PriorityDistanceSquared=StandaloneWorldMesher::TileDistanceSquaredChunks(
Job.Key,NewDesired.CenterChunk);
const int32 BandMin=World->StrateContentCutMinLevel<9&&NewDesired.bHasStrate
? NewDesired.StrateBottomChunkZ*CHUNK_SIZE:MIN_int32;
@@ -167,16 +193,12 @@ public:
for(const FVoxelTileKey& Key:NewDesired.Tiles)
{
if(Resident.contains(Key)||Outstanding.contains(Key))continue;
Queue.push_back({Key,BandMin,BandMax,CrossingIndex});
Queue.push_back({Key,BandMin,BandMax,CrossingIndex,
StandaloneWorldMesher::TileDistanceSquaredChunks(Key,NewDesired.CenterChunk)});
Outstanding.insert(Key);++Stats.NewJobs;
}
CurrentDesired=std::move(NewDesired);
std::sort(Queue.begin(),Queue.end(),[&](const FStandaloneTileJob& A,const FStandaloneTileJob& B)
{
const double DA=StandaloneWorldMesher::TileDistanceSquaredChunks(A.Key,CurrentDesired.CenterChunk);
const double DB=StandaloneWorldMesher::TileDistanceSquaredChunks(B.Key,CurrentDesired.CenterChunk);
return DA==DB?A.Key.Level<B.Key.Level:DA<DB;
});
std::make_heap(Queue.begin(),Queue.end(),FStandaloneTileJobHeapOrder{});
Wake.notify_all();
return Stats;
}
@@ -215,8 +237,21 @@ private:
const int32 Cells=Level<World->FullResClipLevels?CHUNK_SIZE:FMath::Max(1,World->CoarseTileCells);
const int32 Step=FMath::Max(1,Extent/Cells);
int64 SampleCount=0;
const bool bProfile=VoxelDensityProfile::GetMode()!=VoxelDensityProfile::EMode::Disabled;
const VoxelDensityProfile::FSnapshot Before=bProfile
? VoxelDensityProfile::SnapshotCurrentThread():VoxelDensityProfile::FSnapshot{};
FVoxelMeshData Mesh=Mesher.GenerateMesh(Job.Key.OriginVoxels(),Step,Cells,nullptr,
Job.BandZMin,Job.BandZMax,&SampleCount);
if(bProfile)
{
const VoxelDensityProfile::FSnapshot After=VoxelDensityProfile::SnapshotCurrentThread();
for(size_t I=0;I<Out.ProfileCycles.size();++I)
{
const size_t Bucket=I;
Out.ProfileCycles[I]=After.Cycles[Bucket]>=Before.Cycles[Bucket]
?After.Cycles[Bucket]-Before.Cycles[Bucket]:0;
}
}
Out.DensitySamples=static_cast<uint64>(FMath::Max<int64>(0,SampleCount));
Out.Vertices.reserve(static_cast<size_t>(Mesh.Vertices.Num()));
for(int32 I=0;I<Mesh.Vertices.Num();++I)
@@ -241,7 +276,8 @@ private:
std::unique_lock<std::mutex> Lock(Mutex);
Wake.wait(Lock,[&](){return bStop||!Queue.empty();});
if(bStop)return;
Job=Queue.front();Queue.erase(Queue.begin());
std::pop_heap(Queue.begin(),Queue.end(),FStandaloneTileJobHeapOrder{});
Job=Queue.back();Queue.pop_back();
}
FStandaloneCpuTile Tile=Generate(Job);
{