- The number of samples returned is the square number closer to the requested number.
- Updated code paths using the function to request a number of samples close to what they previously got.
#rb Krzysztof.Narkowicz
#preflight 628de9e7f622d972b5732cbb
#ROBOMERGE-AUTHOR: tiago.costa
#ROBOMERGE-SOURCE: CL 20361675 via CL 20361695 via CL 20361702
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v948-20297126)
[CL 20362041 by tiago costa in ue5-main branch]
#ROBOMERGE-AUTHOR: krzysztof.narkowicz
#ROBOMERGE-SOURCE: CL 20082848 via CL 20083225 via CL 20083623
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v943-19904690)
[CL 20106592 by krzysztof narkowicz in ue5-main branch]
* Mostly two sided meshes are assumed to be foliage and forced to be repsented as a cube projection
* Lowered card min density treshold for cube projections
[FYI] Daniel.Wright
#ROBOMERGE-AUTHOR: krzysztof.narkowicz
#ROBOMERGE-SOURCE: CL 19424752 via CL 19424760
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v930-19419903)
[CL 19425489 by krzysztof narkowicz in ue5-main branch]
* Added r.MeshCardRepresentation.Debug for surface cache debugging. It skips DDC reads and writes, allows to regenerate per static mesh on demand and appends extra debug visualization data to every generated mapping
* Surfels are now weighted by opacity - ratio of rays which hit in a given cell, and by visibility - how many rays can hit a given Surfel from outside the mesh. This helps to select most important surfaces to cover when hitting limits
* Removed distance constraint. It�s no longer needed when card discontinuities are already handled inside the surface cache sampling code. This allows to generate larger and more optimal cards
* Limit max number of surfels to prevent generation time from exploding, as dense two sided meshes like large trees can generate many more surfels than simple walls
[FYI] Daniel.Wright
#ROBOMERGE-AUTHOR: krzysztof.narkowicz
#ROBOMERGE-SOURCE: CL 19304117 via CL 19304128
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v926-19321884)
[CL 19346860 by krzysztof narkowicz in ue5-main branch]
Instead of clearing RadiosityAtlas and skipping Radiosity when switching Lumen to SWRT, voxelize the direct lighting that's been calculated so far and keep the existing Radiosity atlas
#preflight 621f057e257fd6e099457e56
#jira UE-144367
#rb Krzysztof.Narkowicz
#lockdown Juan.Canada
#ROBOMERGE-AUTHOR: daniel.wright
#ROBOMERGE-SOURCE: CL 19219209 in //UE5/Release-5.0/... via CL 19219356
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v921-19075845)
[CL 19227351 by daniel wright in ue5-main branch]
#jira UE-122078
#rb Andrew.Davidson, Colin.McGinley
#preflight standard build
#ROBOMERGE-AUTHOR: fred.kimberley
#ROBOMERGE-SOURCE: CL 18817999 in //UE5/Release-5.0/... via CL 18818012 via CL 18822871
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v910-18824042)
[CL 18824721 by fred kimberley in ue5-main branch]
#preflight 61f277cbaf74b68515c0bd64
#ROBOMERGE-AUTHOR: andrew.davidson
#ROBOMERGE-SOURCE: CL 18751911 in //UE5/Release-5.0/... via CL 18751924 via CL 18752067
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v903-18687472)
[CL 18752076 by andrew davidson in ue5-main branch]
* First mesh is voxelized by tracing rays in every voxel
* Then voxel faces are clustered in parallel to build continuos cards. Clustering follows voxel distance, card ratio, density and other metrics
* Finally a specificied number of cards is selected per mesh (user parameter). This allows to tradeoff between coverage and runtime overhead
Other:
* Added surface cache coverage view mode for debugging coverage issues - r.Lumen.Visualize.Mode 6
* Fixed r.LumenScene.SurfaceCache.CardMinResolution to be able to go below 2
#preflight 61d478acdb0309127dfaf8ec
#rb Daniel.Wright
#ROBOMERGE-AUTHOR: krzysztof.narkowicz
#ROBOMERGE-SOURCE: CL 18519098 in //UE5/Release-5.0/... via CL 18519102
#ROBOMERGE-BOT: STARSHIP (Release-Engine-Staging -> Release-Engine-Test) (v899-18417669)
[CL 18519115 by krzysztof narkowicz in ue5-release-engine-test branch]
#ROBOMERGE-SOURCE: CL 16900162 via CL 16917756
#ROBOMERGE-BOT: STARSHIP (Main -> Release-Engine-Test) (v836-16769935)
[CL 16917910 by krzysztof narkowicz in ue5-release-engine-test branch]
* Generate surfels directly from the triangles instead of relying on ray tracing, to fix coverage issues on some meshes
* Surfels which are inside meshes (surrounded by back faces) or are too close to geometry are discarded
* New surfel clustering algorithm, which inserts one seed after another and tries to iteratively grow clusters in order to find the best set of seeds. Final step is to reset all clusters and grow all simultaneously from previously selected seeds.
* Cluster growing is based on normal, distance and surfel visibility (don�t cluster surfels near geometry first, as it can cause algorithm to be stuck in a local minimum)
* Runtime sampling has strict culling based on the angle and card AABB. Additionally, the tri-planar blending zone was tightened. This improves performance. 0.62-0.5 ms.
* Added various visualizations and CVars for card generation debugging
* Fixed card visibility bug, where card could influence outside of it�s range due to negative shadow map visibility
[FYI] Daniel.Wright, Patrick.Kelly
#ROBOMERGE-SOURCE: CL 16897632 via CL 16916104
#ROBOMERGE-BOT: STARSHIP (Main -> Release-Engine-Test) (v836-16769935)
[CL 16916378 by krzysztof narkowicz in ue5-release-engine-test branch]
#ROBOMERGE-SOURCE: CL 15788010 in //UE5/Release-5.0-EarlyAccess/...
#ROBOMERGE-BOT: STARSHIP (Release-5.0-EarlyAccess -> Main) (v783-15756269)
[CL 15790801 by daniel wright in ue5-main branch]
* SDFs are now generated, allocated from the atlas and uploaded in 8^3 bricks (7^3 unique data, half voxel padding).
* Tracing must load the brick index from the indirection table, and only bricks near the surface are stored
* 3 mips are now generated, with the lowest resolution always loaded and the other 2 streamed
* SDFs are now G8 narrow band. Lower resolution mips must be traversed when querying distance to nearest surface far away from the surface
* The Distance Field Brick Atlas is now stored for each FScene and dynamically resized based on needs with a GPU memcopy
* Brick atlas uses a 1d pooled allocator which has no fragmentation and greatly reduces packing waste over the 3d allocator
* Added new indirection for Distance Field Asset data, so that only a single entry needs to be updated when a mip is streamed in or out in scenes with millions of instances
* Compute shaders operating on distance field instances generate streaming requests, which are async read back to CPU, turned into IO requests, which are polled and when complete uploaded to atlases
* Any mesh instance inside the Global SDF extent (200m) requests mip1, and at 50m requests mip2
* Now using a batched compute scatter to upload to the distance field atlas instead of RHIUpdateTexture3d, to bypass alignment restrictions and per-upload overhead
* Distance Field streaming uses an async task to move Memcpy and IO request overhead off of the Rendering Thread
* Distance Field Visualization now computes a normal from the SDF gradient and does simple lighting to better visualize the scene representation
* Increased r.DistanceFields.MaxPerMeshResolution from 128 to 512, to better represent large objects
* Mesh SDF generation now uses an Embree point query to calculate closest unsigned distance, and then a much smaller set of rays to count backfaces for negative region determination, for a 11x speedup
* Upgraded mesh utilities to Embree 3.12.2 to get point queries
* Fixed wrong transform used for SDF normals in Lumen, causing non-uniformly scaled meshes to have incorrect Surface Cache interpolation
* Fixed Static Mesh materials not getting PostLoaded before SDF build, causing their blend modes to be wrong for the build, which corrupts the DDC. Also included those blend modes in the DDC key.
Original costs on 1080 GTX (full updates on everything and no screen traces)
10.60ms UpdateGlobalDistanceField
3.62ms LumenReflectiveTest.DirectionalLight_1 Shadowmap 1
1.73ms VoxelizeCards Clipmaps=[0,1,2,3]
0.38ms TraceCards 1 dispatch 1 groups
0.51ms TraceCards 1 dispatch 1 groups
Sparse SDF costs
12.06ms UpdateGlobalDistanceField
4.35ms LumenReflectiveTest.DirectionalLight_1 Shadowmap 1
2.30ms VoxelizeCards Clipmaps=[0,1,2,3]
0.69ms TraceCards 1 dispatch 1 groups
0.77ms TraceCards 1 dispatch 1 groups
Tested: TopazEntry PC, Reverb PC and PS5, EngineTests, QAGame, Rift, Frosty P_Construct_WP, FortGPUTestbed
#rb Krzysztof.Narkowicz
#ROBOMERGE-OWNER: Daniel.Wright
#ROBOMERGE-AUTHOR: daniel.wright
#ROBOMERGE-SOURCE: CL 15784493 in //UE5/Release-5.0-EarlyAccess/...
#ROBOMERGE-BOT: STARSHIP (Release-5.0-EarlyAccess -> Main) (v783-15756269)
#ROBOMERGE-CONFLICT from-shelf
[CL 15790658 by Daniel Wright in ue5-main branch]
* Moved LumenScene functionality into new LumenScene.cpp file
* Moved all Lumen data into FLumenPrimitive. FPrimitiveSceneInfo now contains just a LumenPrimitiveIndex
* Mesh cards are spawned only near the camera on demand. This saves ~200-300mb of GPU memory on big city levels, and also skips some GPU and CPU costs
* Fixed card merging (previously it was reusing MeshCards which could result in missing cards)
#ROBOMERGE-SOURCE: CL 15494971 in //UE5/Release-5.0-EarlyAccess/...
#ROBOMERGE-BOT: STARSHIP (Release-5.0-EarlyAccess -> Main) (v771-15082668)
[CL 15494982 by krzysztof narkowicz in ue5-main branch]
* r.LumenScene.MeshCardsMaxLOD controls representation quality. With 0 being a box projection
* New card placement to focus on important parts of a mesh (skip non visible parts)
* Replaced CubeMapTrees with MeshCards. Instead of a big volume texture lookup there are now 6 lists of decals (one per axis)
* Merged distance field and mesh card Embree setup code into a set of common functions
* Removed hardcoded limits of 6 cards per mesh
#fyi Daniel.Wright, Patrick.Kelly, Yuriy.ODonnell
[CL 15144947 by Krzysztof Narkowicz in ue5-main branch]
This is later used to build static mesh derived data in cases where Nanite coarse mesh representation replaces original static mesh data and we can't depend on it. In extreme cases it lowers main thread derived data build time from ~7.5s to ~0.1s.
#rb Richard.TalbotWatkin
[CL 14578414 by Krzysztof Narkowicz in ue5-main branch]
