This represents UE4/Main @17911760, Release-5.0 @17915875 and Dev-PerfTest @17914035
[CL 17918595 by aurel cordonnier in ue5-release-engine-test branch]
This represents UE4/Main @17774255, Release-5.0 @17791557 and Dev-PerfTest @17789485
[CL 17794212 by aurel cordonnier in ue5-release-engine-test branch]
- Cook side extra low lod coarse mesh will be created at 1/8th of the original coarse mesh if mesh streaming or editor data is available for the platform
- Nanite static meshes with streamable LODs are registered to a Nanite::CoarseMeshStreamingManager which will handle the stream in/out requests for these assets in fixed memory budget (memory budget is part of the total mesh streaming budget)
- Nanite proxy needs to handle different raytracing proxy LOD indices now for dynamic and cached indices
- What to stream is driven by what's used in the TLAS build - each deferred render pass all used streamable meshes are collected and forwarded to the CoarseMeshStreamingManager
- CoarseMeshStreamingManager is updated after all views are rendered and will make stream in/out requests depending on the TLAS usage
- When LODs of UStaticMeshes renderdata is loaded or unloaded then a rebuild of the cached render data is requested to make sure the correct LOD & RT BLAS is used (to do this the UStaticMeshComponent are registered to the UStaticMesh so a request for rebuild can be made for all the proxies on the scene)
#preflight 612659e872e9eb00010a27d9
#ROBOMERGE-SOURCE: CL 17314263 via CL 17315687
#ROBOMERGE-BOT: STARSHIP (Main -> Release-Engine-Test) (v861-17282326)
[CL 17315942 by kenzo terelst in ue5-release-engine-test branch]
This involves a change to the DDC key format for static meshes with multiple sections in LOD0.
#jira UE-112945
#rb Graham.Wihlidal, Alexis.Matte
[CL 16127723 by Richard TalbotWatkin in ue5-main branch]
This CL also updates the bounding box computation, previously the bounds were computed on LOD0 render data, ie after the nanite cut. Now the bounds of either the HiRes SourceModel, or high-res LOD0, are combined with the final LOD0 bounds.
#rb brian.karis, richard.talbotwatkin, jimmy.andrews, david.hill
#rnx
#jira none
#ROBOMERGE-SOURCE: CL 15485739 in //UE5/Release-5.0-EarlyAccess/...
#ROBOMERGE-BOT: STARSHIP (Release-5.0-EarlyAccess -> Main) (v771-15082668)
[CL 15485742 by ryan schmidt in ue5-main branch]
Phase 1 adds the async compilation without introducing any higher-level awareness.
This step provides the biggest performance improvements by allowing skeletal mesh to be compiled in
multiple threads but stalls are expected to occur before the level has finished loading as soon as
properties of skeletal meshes are accessed by the game-thread.
- Make skeletal mesh reduction and renderdata building thread-safe where needed
- Add protection on all USkeletalMesh accessors so the game-thread can wait on data still being built if needed
- Modify both PostLoad and Build to be async when the feature is enabled
- Make FRawSkeletalMeshBulkData::LoadRawMesh thread-safe
- Fix some PropertyChangedEvent name comparison to use strongly typed GET_MEMBER_NAME_CHECKED
- Add an experimental setting options to enable the feature (disabled by default for now)
DEBUGGING
- Can be forcibly enabled/disabled through command-line via -asyncskeletalmeshcompilation=[off, on, paused]
- Can pause skeletal mesh compilation using Editor.AsyncSkeletalMeshCompilation = 2 or -asyncskeletalmeshcompilation=paused
- Can manually resume a specified amount of paused compilation using Editor.AsyncSkeletalMeshCompilationResume [Num]
- Can forcibly wait on all compilation using Editor.AsyncSkeletalMeshCompilationFlushAll
BENCHMARKS
Tested on AMD TR 3970X with 256GB RAM
- 6m50s to 2m58s for loading P_Construct with a local-only DDC that doesn't contains any prebuilt skeletal mesh
- 3m10s to 39s for importing 299 SK with the Interchange framework
TESTS
- Opened all 5769 SK from FortniteGame to exercise all legacy conversion code paths
- Opened all 102 SK from QAGame
- Opened P_Construct, Apollo_Terrain, LumenTest, P_World
- Imported 299 SK with and without Interchange activated
- Ensure no new regression introduced in any of the 70 SK related tests from EngineTest
- Run all FBX import test in EngineTest
- Cook / Run ShooterGame Client
#rb Alexis.Matte
[CL 15452895 by danny couture in ue5-main branch]
Common mistakes:
a) TRACE_CPUPROFILER_EVENT_SCOPE("Foo") or TRACE_CPUPROFILER_EVENT_SCOPE("FClass::Foo")
--> results in a timer named "Foo" or "FClass::Foo" (the quotes will be included in the name)
b) TRACE_CPUPROFILER_EVENT_SCOPE(TEXT("Foo"))
--> results in a timer named L"Foo" (L and quotes will be included in the name)
c) TRACE_CPUPROFILER_EVENT_SCOPE_TEXT(TEXT("Other Foo")) or TRACE_CPUPROFILER_EVENT_SCOPE_TEXT("Foo")
--> Slow! It will use dynamic string matching that adds an unnecessary overhead.
Correct usage:
TRACE_CPUPROFILER_EVENT_SCOPE(Foo)
TRACE_CPUPROFILER_EVENT_SCOPE(FClass::Foo)
TRACE_CPUPROFILER_EVENT_SCOPE_STR("Other Foo") // when timer name has spaces
TRACE_CPUPROFILER_EVENT_SCOPE_TEXT(*Foo.ToString()) // only if a dynamic name is really needed
#rb Catalin.Dragoiu
#fyi Marc.Audy, Krzysztof.Narkowicz, Rune.Stubbe, Michal.Valient
[CL 15134822 by ionut matasaru in ue5-main branch]
Adding a public member to USkeletalMesh class is now prohibed since it can break asynchronous operations that will be added next to this submit.
#rb danny.couture
#jira UEENT-3936
#rnx
[CL 14812920 by Alexis Matte in ue5-main branch]
