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

wip(standalone): GPU step 7 partial -- parity check deferred off cold path (stopped early)

Browse Source 
Luna measured the startup density-parity check at 1,049 ms, nearly all of the
~1.13 s before shape compile; it now runs after first display and still passes
3,436/3,436. CPU hashes 44/44 at 1 and 8 workers, walkability 5,642/5,642,
headless 8-worker clipmap 0.769 s. Stopped by the owner mid-round, while
auditing the remote far holes; first-launch stall and hybrid speed untouched.

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 2026-10-06 20:17:37 +02:00
1 parent b0a14262b1
commit 41716f00b8
6 files changed
+198 -24
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

Standalone/Source/StandaloneViewer.cpp
+50 -4
View File
@@ -1491,6 +1491,7 @@ namespace
std::filesystem::path GpuFarShaderPath;
double GpuFarEnqueueWallSeconds=0.0;
uint64 GpuFarTilesSubmitted=0;
uint64 PresentCallCount=0;
double GpuFarGpuDispatchMs=0.0;
double GpuFarGpuDispatchMaxMs=0.0;
double GpuFarSubmissionSpanMs=0.0,GpuFarSubmissionSpanMaxMs=0.0;
@@ -1823,6 +1824,7 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
size_t PrintParity()
{
if(bParityChecked||!World)return 0;
const auto ParityStart=std::chrono::steady_clock::now();
bParityChecked=true;size_t MismatchCount=0;
std::map<std::string,size_t> Kinds;
for(size_t I=0;I<World->DensitySamples.size();++I)
@@ -1841,18 +1843,28 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
std::printf("Density parity: %zu/%zu bit-identical; mismatches=%zu\n",
World->DensitySamples.size()-MismatchCount,World->DensitySamples.size(),MismatchCount);
for(const auto& [Kind,Count]:Kinds)std::printf(" %-24s %zu\n",Kind.c_str(),Count);
std::fflush(stdout);
std::printf("[DiagnosticProfile] stage=density_parity samples=%zu duration_ms=%.3f\n",
World->DensitySamples.size(),std::chrono::duration<double,std::milli>(
std::chrono::steady_clock::now()-ParityStart).count());
std::fflush(stdout);
return MismatchCount;
}
bool ConfigureWorld(std::string& Error)
{
const auto ConfigureStart=std::chrono::steady_clock::now();
ShapeWorld.reset();
Snapshot={};
World=std::make_shared<FStandaloneOwnerWorld>();
const auto ResolveStart=std::chrono::steady_clock::now();
if(!StandaloneOwnerLoader::Resolve(OwnerJsonRoot,*World,Error))return false;
if(!bParityChecked&&Seed==World->Seed&&!bRoundTunnels)PrintParity();
const auto ResolveEnd=std::chrono::steady_clock::now();
std::printf("[ConfigureProfile] stage=owner_resolve duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(ResolveEnd-ResolveStart).count());
if(!bParityChecked&&Seed==World->Seed&&!bRoundTunnels)
std::printf("[DiagnosticDeferred] density_parity=--parity-only samples=%zu\n",
World->DensitySamples.size());
const auto ClipOverrideStart=std::chrono::steady_clock::now();
const int32 ExportClipRadius=World->ClipRadius;
World->ClipRadius=ClipRadius;
World->Settings->ClipRadius=ClipRadius;
@@ -1864,8 +1876,13 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
std::printf("Standalone clip override: exported radius=%d (112 m cube) -> radius=%d (%.0f m LOD0 cube), voxel size=25 cm.\n",
ExportClipRadius,ClipRadius,FullDetailWidth);
}
std::printf("[ConfigureProfile] stage=clip_override duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(std::chrono::steady_clock::now()-ClipOverrideStart).count());
const auto DensitySetupStart=std::chrono::steady_clock::now();
if(!PrepareStandaloneDensityWorld(*World,Seed,bRoundTunnels,
bRoundTunnelEnhanced,Error))return false;
std::printf("[ConfigureProfile] stage=density_world_setup duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(std::chrono::steady_clock::now()-DensitySetupStart).count());
if(MaxClipLevelOverride>=0)
{
World->MaxClipLevel=MaxClipLevelOverride;
@@ -1875,14 +1892,24 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
}
// Compile one world field before streaming. The streamer already prioritizes the
// player's neighborhood, so a second near-only field would only duplicate work.
if(bShapeWorld&&!BuildShapeWorldForCamera(Error,true,false))return false;
const auto ShapeSetupStart=std::chrono::steady_clock::now();
if(bShapeWorld&&!BuildShapeWorldForCamera(Error,true,false,true))return false;
std::printf("[ConfigureProfile] stage=shape_world_request duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(std::chrono::steady_clock::now()-ShapeSetupStart).count());
const auto CacheStart=std::chrono::steady_clock::now();
LoadMeshCache();
std::printf("[ConfigureProfile] stage=mesh_cache duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(std::chrono::steady_clock::now()-CacheStart).count());
std::printf("[ConfigureProfile] stage=configure_total duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(std::chrono::steady_clock::now()-ConfigureStart).count());
return true;
}
bool BuildShapeWorldForCamera(std::string& Error,bool bIncludeFarWorld=false,
bool bNeighborhoodOnly=false)
bool bNeighborhoodOnly=false,
bool bDeferWalkabilityValidation=false)
{
const auto RequestSetupStart=std::chrono::steady_clock::now();
const FVector CameraVoxels(Position.X*4.0,Position.Y*4.0,Position.Z*4.0);
const FStandaloneDesiredTileSet Desired=StandaloneWorldMesher::BuildDesiredSet(*World,CameraVoxels);
FIntVector Min(INT32_MAX,INT32_MAX,INT32_MAX),Max(INT32_MIN,INT32_MIN,INT32_MIN);
@@ -1917,6 +1944,11 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
Min.X-=NearBuildPad;Min.Y-=NearBuildPad;Min.Z-=NearBuildPad;
Max.X+=NearBuildPad;Max.Y+=NearBuildPad;Max.Z+=NearBuildPad;
}
const auto DesiredSetupEnd=std::chrono::steady_clock::now();
std::printf("[ConfigureProfile] stage=desired_tiles_and_bounds tiles=%llu duration_ms=%.3f\n",
static_cast<unsigned long long>(FineTiles),std::chrono::duration<double,std::milli>(
DesiredSetupEnd-RequestSetupStart).count());
const auto SnapshotStart=std::chrono::steady_clock::now();
Snapshot=VoxelForgeCore::MakeStandaloneSnapshotFromParsedJson(OwnerJsonRoot,Seed,ClipRadius,
MaxClipLevelOverride,bRoundTunnels,bRoundTunnelEnhanced,Error);
if(!Snapshot)
@@ -1931,7 +1963,12 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
Request.includeFarWorld=bIncludeFarWorld;Request.roundTunnels=bRoundTunnels;
Request.enhancedTunnelWalls=bRoundTunnelEnhanced;
Request.profileDetailBand=bProfileDetailBand;
Request.validatePassageRoutes=!bDeferWalkabilityValidation;
if(bDeferWalkabilityValidation)
std::printf("[DiagnosticDeferred] walkability=full_headless_or_route_audit workers=8 samples=5642\n");
const auto ShapeCompileStart=std::chrono::steady_clock::now();
std::printf("[ConfigureProfile] stage=snapshot_and_build_request duration_ms=%.3f\n",
std::chrono::duration<double,std::milli>(ShapeCompileStart-SnapshotStart).count());
std::printf("[ColdTimeline] stage=shape_compile begin_ms=%.3f\n",
std::chrono::duration<double,std::milli>(ShapeCompileStart-ColdProcessStart).count());
if(!ShapeWorld->compile(Request,CoreScheduler,Error))
@@ -1980,6 +2017,7 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
Request.includeFarWorld=true;Request.roundTunnels=bRoundTunnels;
Request.enhancedTunnelWalls=bRoundTunnelEnhanced;
Request.profileDetailBand=bProfileDetailBand;
Request.validatePassageRoutes=false;
bFullShapeBuildSucceeded=PendingFullShapeWorld->compile(
Request,CoreScheduler,FullShapeBuildError);
bFullShapeBuildFinished.store(true,std::memory_order_release);
@@ -3357,6 +3395,14 @@ float3 Lit=Rock*(Ambient+0.88*Diffuse)+float3(0.50,0.69,0.82)*Head*0.75;
const auto PresentCallStart=std::chrono::steady_clock::now();
SwapChain->Present(PresentSyncInterval,0);
const auto PresentCallEnd=std::chrono::steady_clock::now();
const uint64 PresentOrdinal=++PresentCallCount;
if(PresentOrdinal<=3)
std::printf("[PresentProfile] ordinal=%llu at_ms=%.3f sync=%u call_ms=%.3f gpu_tiles_submitted=%llu gpu_batches=%llu\n",
static_cast<unsigned long long>(PresentOrdinal),
std::chrono::duration<double,std::milli>(PresentCallEnd-ColdProcessStart).count(),
PresentSyncInterval,std::chrono::duration<double,std::milli>(PresentCallEnd-PresentCallStart).count(),
static_cast<unsigned long long>(GpuFarTilesSubmitted),
static_cast<unsigned long long>(GpuFarSubmissionBatches));
UpdateTitle();
const auto DrawEnd=std::chrono::steady_clock::now();
const double CpuPumpMs=std::chrono::duration<double,std::milli>(CpuPumpEnd-DrawStart).count();