bevy/crates/bevy_pbr/src/meshlet/meshlet_mesh_manager.rs

163 lines
6.4 KiB
Rust
Raw Normal View History

Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
use super::{
asset::{Meshlet, MeshletBoundingSpheres, MeshletSimplificationError},
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
persistent_buffer::PersistentGpuBuffer,
MeshletMesh,
};
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
use alloc::sync::Arc;
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
use bevy_asset::{AssetId, Assets};
use bevy_ecs::{
system::{Res, ResMut, Resource},
world::{FromWorld, World},
};
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
use bevy_math::Vec2;
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
use bevy_render::{
render_resource::BufferAddress,
renderer::{RenderDevice, RenderQueue},
};
use bevy_utils::HashMap;
Add `core` and `alloc` over `std` Lints (#15281) # Objective - Fixes #6370 - Closes #6581 ## Solution - Added the following lints to the workspace: - `std_instead_of_core` - `std_instead_of_alloc` - `alloc_instead_of_core` - Used `cargo +nightly fmt` with [item level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Item%5C%3A) to split all `use` statements into single items. - Used `cargo clippy --workspace --all-targets --all-features --fix --allow-dirty` to _attempt_ to resolve the new linting issues, and intervened where the lint was unable to resolve the issue automatically (usually due to needing an `extern crate alloc;` statement in a crate root). - Manually removed certain uses of `std` where negative feature gating prevented `--all-features` from finding the offending uses. - Used `cargo +nightly fmt` with [crate level use formatting](https://rust-lang.github.io/rustfmt/?version=v1.6.0&search=#Crate%5C%3A) to re-merge all `use` statements matching Bevy's previous styling. - Manually fixed cases where the `fmt` tool could not re-merge `use` statements due to conditional compilation attributes. ## Testing - Ran CI locally ## Migration Guide The MSRV is now 1.81. Please update to this version or higher. ## Notes - This is a _massive_ change to try and push through, which is why I've outlined the semi-automatic steps I used to create this PR, in case this fails and someone else tries again in the future. - Making this change has no impact on user code, but does mean Bevy contributors will be warned to use `core` and `alloc` instead of `std` where possible. - This lint is a critical first step towards investigating `no_std` options for Bevy. --------- Co-authored-by: François Mockers <francois.mockers@vleue.com>
2024-09-27 00:59:59 +00:00
use core::ops::Range;
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
/// Manages uploading [`MeshletMesh`] asset data to the GPU.
#[derive(Resource)]
pub struct MeshletMeshManager {
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
pub vertex_positions: PersistentGpuBuffer<Arc<[u32]>>,
pub vertex_normals: PersistentGpuBuffer<Arc<[u32]>>,
pub vertex_uvs: PersistentGpuBuffer<Arc<[Vec2]>>,
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
pub indices: PersistentGpuBuffer<Arc<[u8]>>,
pub meshlets: PersistentGpuBuffer<Arc<[Meshlet]>>,
pub meshlet_bounding_spheres: PersistentGpuBuffer<Arc<[MeshletBoundingSpheres]>>,
pub meshlet_simplification_errors: PersistentGpuBuffer<Arc<[MeshletSimplificationError]>>,
meshlet_mesh_slices: HashMap<AssetId<MeshletMesh>, [Range<BufferAddress>; 7]>,
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
}
impl FromWorld for MeshletMeshManager {
fn from_world(world: &mut World) -> Self {
let render_device = world.resource::<RenderDevice>();
Self {
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
vertex_positions: PersistentGpuBuffer::new("meshlet_vertex_positions", render_device),
vertex_normals: PersistentGpuBuffer::new("meshlet_vertex_normals", render_device),
vertex_uvs: PersistentGpuBuffer::new("meshlet_vertex_uvs", render_device),
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
indices: PersistentGpuBuffer::new("meshlet_indices", render_device),
meshlets: PersistentGpuBuffer::new("meshlets", render_device),
meshlet_bounding_spheres: PersistentGpuBuffer::new(
"meshlet_bounding_spheres",
render_device,
),
meshlet_simplification_errors: PersistentGpuBuffer::new(
"meshlet_simplification_errors",
render_device,
),
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
meshlet_mesh_slices: HashMap::new(),
}
}
}
impl MeshletMeshManager {
pub fn queue_upload_if_needed(
&mut self,
asset_id: AssetId<MeshletMesh>,
assets: &mut Assets<MeshletMesh>,
) -> Range<u32> {
let queue_meshlet_mesh = |asset_id: &AssetId<MeshletMesh>| {
let meshlet_mesh = assets.remove_untracked(*asset_id).expect(
"MeshletMesh asset was already unloaded but is not registered with MeshletMeshManager",
);
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
let vertex_positions_slice = self
.vertex_positions
.queue_write(Arc::clone(&meshlet_mesh.vertex_positions), ());
let vertex_normals_slice = self
.vertex_normals
.queue_write(Arc::clone(&meshlet_mesh.vertex_normals), ());
let vertex_uvs_slice = self
.vertex_uvs
.queue_write(Arc::clone(&meshlet_mesh.vertex_uvs), ());
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
let indices_slice = self
.indices
.queue_write(Arc::clone(&meshlet_mesh.indices), ());
let meshlets_slice = self.meshlets.queue_write(
Arc::clone(&meshlet_mesh.meshlets),
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
(
vertex_positions_slice.start,
vertex_normals_slice.start,
indices_slice.start,
),
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
);
let meshlet_bounding_spheres_slice = self
.meshlet_bounding_spheres
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
.queue_write(Arc::clone(&meshlet_mesh.meshlet_bounding_spheres), ());
let meshlet_simplification_errors_slice = self
.meshlet_simplification_errors
.queue_write(Arc::clone(&meshlet_mesh.meshlet_simplification_errors), ());
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
[
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
vertex_positions_slice,
vertex_normals_slice,
vertex_uvs_slice,
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
indices_slice,
meshlets_slice,
meshlet_bounding_spheres_slice,
meshlet_simplification_errors_slice,
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
]
};
// If the MeshletMesh asset has not been uploaded to the GPU yet, queue it for uploading
let [_, _, _, _, meshlets_slice, _, _] = self
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
.meshlet_mesh_slices
.entry(asset_id)
.or_insert_with_key(queue_meshlet_mesh)
.clone();
let meshlets_slice_start = meshlets_slice.start as u32 / size_of::<Meshlet>() as u32;
let meshlets_slice_end = meshlets_slice.end as u32 / size_of::<Meshlet>() as u32;
meshlets_slice_start..meshlets_slice_end
}
pub fn remove(&mut self, asset_id: &AssetId<MeshletMesh>) {
if let Some(
[vertex_positions_slice, vertex_normals_slice, vertex_uvs_slice, indices_slice, meshlets_slice, meshlet_bounding_spheres_slice, meshlet_simplification_errors_slice],
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
) = self.meshlet_mesh_slices.remove(asset_id)
{
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
self.vertex_positions
.mark_slice_unused(vertex_positions_slice);
self.vertex_normals.mark_slice_unused(vertex_normals_slice);
self.vertex_uvs.mark_slice_unused(vertex_uvs_slice);
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
self.indices.mark_slice_unused(indices_slice);
self.meshlets.mark_slice_unused(meshlets_slice);
self.meshlet_bounding_spheres
.mark_slice_unused(meshlet_bounding_spheres_slice);
self.meshlet_simplification_errors
.mark_slice_unused(meshlet_simplification_errors_slice);
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
}
}
}
/// Upload all newly queued [`MeshletMesh`] asset data to the GPU.
pub fn perform_pending_meshlet_mesh_writes(
mut meshlet_mesh_manager: ResMut<MeshletMeshManager>,
render_queue: Res<RenderQueue>,
render_device: Res<RenderDevice>,
) {
meshlet_mesh_manager
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
.vertex_positions
.perform_writes(&render_queue, &render_device);
meshlet_mesh_manager
.vertex_normals
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
.perform_writes(&render_queue, &render_device);
meshlet_mesh_manager
Per-meshlet compressed vertex data (#15643) # Objective - Prepare for streaming by storing vertex data per-meshlet, rather than per-mesh (this means duplicating vertices per-meshlet) - Compress vertex data to reduce the cost of this ## Solution The important parts are in from_mesh.rs, the changes to the Meshlet type in asset.rs, and the changes in meshlet_bindings.wgsl. Everything else is pretty secondary/boilerplate/straightforward changes. - Positions are quantized in centimeters with a user-provided power of 2 factor (ideally auto-determined, but that's a TODO for the future), encoded as an offset relative to the minimum value within the meshlet, and then stored as a packed list of bits using the minimum number of bits needed for each vertex position channel for that meshlet - E.g. quantize positions (lossly, throws away precision that's not needed leading to using less bits in the bitstream encoding) - Get the min/max quantized value of each X/Y/Z channel of the quantized positions within a meshlet - Encode values relative to the min value of the meshlet. E.g. convert from [min, max] to [0, max - min] - The new max value in the meshlet is (max - min), which only takes N bits, so we only need N bits to store each channel within the meshlet (lossless) - We can store the min value and that it takes N bits per channel in the meshlet metadata, and reconstruct the position from the bitstream - Normals are octahedral encoded and than snorm2x16 packed and stored as a single u32. - Would be better to implement the precise variant of octhedral encoding for extra precision (no extra decode cost), but decided to keep it simple for now and leave that as a followup - Tried doing a quantizing and bitstream encoding scheme like I did for positions, but struggled to get it smaller. Decided to go with this for simplicity for now - UVs are uncompressed and take a full 64bits per vertex which is expensive - In the future this should be improved - Tangents, as of the previous PR, are not explicitly stored and are instead derived from screen space gradients - While I'm here, split up MeshletMeshSaverLoader into two separate types Other future changes include implementing a smaller encoding of triangle data (3 u8 indices = 24 bits per triangle currently), and more disk-oriented compression schemes. References: * "A Deep Dive into UE5's Nanite Virtualized Geometry" https://advances.realtimerendering.com/s2021/Karis_Nanite_SIGGRAPH_Advances_2021_final.pdf#page=128 (also available on youtube) * "Towards Practical Meshlet Compression" https://arxiv.org/pdf/2404.06359 * "Vertex quantization in Omniforce Game Engine" https://daniilvinn.github.io/2024/05/04/omniforce-vertex-quantization.html ## Testing - Did you test these changes? If so, how? - Converted the stanford bunny, and rendered it with a debug material showing normals, and confirmed that it's identical to what's on main. EDIT: See additional testing in the comments below. - Are there any parts that need more testing? - Could use some more size comparisons on various meshes, and testing different quantization factors. Not sure if 4 is a good default. EDIT: See additional testing in the comments below. - Also did not test runtime performance of the shaders. EDIT: See additional testing in the comments below. - How can other people (reviewers) test your changes? Is there anything specific they need to know? - Use my unholy script, replacing the meshlet example https://paste.rs/7xQHk.rs (must make MeshletMesh fields pub instead of pub crate, must add lz4_flex as a dev-dependency) (must compile with meshlet and meshlet_processor features, mesh must have only positions, normals, and UVs, no vertex colors or tangents) --- ## Migration Guide - TBD by JMS55 at the end of the release
2024-10-08 18:42:55 +00:00
.vertex_uvs
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
.perform_writes(&render_queue, &render_device);
meshlet_mesh_manager
.indices
.perform_writes(&render_queue, &render_device);
meshlet_mesh_manager
.meshlets
.perform_writes(&render_queue, &render_device);
meshlet_mesh_manager
.meshlet_bounding_spheres
.perform_writes(&render_queue, &render_device);
meshlet_mesh_manager
.meshlet_simplification_errors
.perform_writes(&render_queue, &render_device);
Meshlet software raster + start of cleanup (#14623) # Objective - Faster meshlet rasterization path for small triangles - Avoid having to allocate and write out a triangle buffer - Refactor gpu_scene.rs ## Solution - Replace the 32bit visbuffer texture with a 64bit visbuffer buffer, where the left 32 bits encode depth, and the right 32 bits encode the existing cluster + triangle IDs. Can't use 64bit textures, wgpu/naga doesn't support atomic ops on textures yet. - Instead of writing out a buffer of packed cluster + triangle IDs (per triangle) to raster, the culling pass now writes out a buffer of just cluster IDs (per cluster, so less memory allocated, cheaper to write out). - Clusters for software raster are allocated from the left side - Clusters for hardware raster are allocated in the same buffer, from the right side - The buffer size is fixed at MeshletPlugin build time, and should be set to a reasonable value for your scene (no warning on overflow, and no good way to determine what value you need outside of renderdoc - I plan to fix this in a future PR adding a meshlet stats overlay) - Currently I don't have a heuristic for software vs hardware raster selection for each cluster. The existing code is just a placeholder. I need to profile on a release scene and come up with a heuristic, probably in a future PR. - The culling shader is getting pretty hard to follow at this point, but I don't want to spend time improving it as the entire shader/pass is getting rewritten/replaced in the near future. - Software raster is a compute workgroup per-cluster. Each workgroup loads and transforms the <=64 vertices of the cluster, and then rasterizes the <=64 triangles of the cluster. - Two variants are implemented: Scanline for clusters with any larger triangles (still smaller than hardware is good at), and brute-force for very very tiny triangles - Once the shader determines that a pixel should be filled in, it does an atomicMax() on the visbuffer to store the results, copying how Nanite works - On devices with a low max workgroups per dispatch limit, an extra compute pass is inserted before software raster to convert from a 1d to 2d dispatch (I don't think 3d would ever be necessary). - I haven't implemented the top-left rule or subpixel precision yet, I'm leaving that for a future PR since I get usable results without it for now - Resources used: https://kristoffer-dyrkorn.github.io/triangle-rasterizer and chapters 6-8 of https://fgiesen.wordpress.com/2013/02/17/optimizing-sw-occlusion-culling-index - Hardware raster now spawns 64*3 vertex invocations per meshlet, instead of the actual meshlet vertex count. Extra invocations just early-exit. - While this is slower than the existing system, hardware draws should be rare now that software raster is usable, and it saves a ton of memory using the unified cluster ID buffer. This would be fixed if wgpu had support for mesh shaders. - Instead of writing to a color+depth attachment, the hardware raster pass also does the same atomic visbuffer writes that software raster uses. - We have to bind a dummy render target anyways, as wgpu doesn't currently support render passes without any attachments - Material IDs are no longer written out during the main rasterization passes. - If we had async compute queues, we could overlap the software and hardware raster passes. - New material and depth resolve passes run at the end of the visbuffer node, and write out view depth and material ID depth textures ### Misc changes - Fixed cluster culling importing, but never actually using the previous view uniforms when doing occlusion culling - Fixed incorrectly adding the LOD error twice when building the meshlet mesh - Splitup gpu_scene module into meshlet_mesh_manager, instance_manager, and resource_manager - resource_manager is still too complex and inefficient (extract and prepare are way too expensive). I plan on improving this in a future PR, but for now ResourceManager is mostly a 1:1 port of the leftover MeshletGpuScene bits. - Material draw passes have been renamed to the more accurate material shade pass, as well as some other misc renaming (in the future, these will be compute shaders even, and not actual draw calls) --- ## Migration Guide - TBD (ask me at the end of the release for meshlet changes as a whole) --------- Co-authored-by: vero <email@atlasdostal.com>
2024-08-26 17:54:34 +00:00
}