Commit Graph

45 Commits

Author SHA1 Message Date
Rémi Verschelde
2a8a59eac0
Merge pull request #50644 from BastiaanOlij/mobile_compute_to_fragment 2021-07-25 11:58:11 +02:00
Bastiaan Olij
1f69582835 Use fragment shader instead of compute shader for effects for mobile renderer 2021-07-25 13:58:21 +10:00
reduz
9293bc3935 Implement more rendering options as specialization constants
* Shadow quality settings now specialization constant.
* Decal and light projector filters can be set.
* Changing those settings forces re-creation of the pipelines.

These changes should help improve performance related to shadow mapping, and allows improving performance by sacrificing decal and light projector quality.
2021-07-19 21:51:29 -03:00
reduz
76e25438c9 Fixes to forward mobile
* use valid format for framebuffer: VK_FORMAT_A2B10G10R10_UNORM_PACK32
* Unfortunately cant be used for compute.
* Mobile will need to do refprobe, sky, mipmapblurring using raster.
2021-07-15 21:05:58 -03:00
Juan Linietsky
3f12765a7e
Merge pull request #49678 from BastiaanOlij/mobile_renderer_select_color_format
Use 32bit instead of 64bit 3D render buffer on mobile renderer
2021-07-15 10:16:39 -03:00
Bastiaan Olij
8a29208fc4 Use 32bit instead of 64bit 3D render buffer on mobile 2021-07-13 10:26:28 +10:00
reduz
ad9f606ed8 Use specialization constants in clustered renderer
* Keep track of when projector, softshadow or directional sofshadow were enabled.
* Enable them via specializaton constant where it makes sense.
* Re-implements soft shadows.
* Re-implements light projectors.
2021-07-12 20:33:52 -03:00
reduz
6c55d2aad2 Fix Render Info
* Fixed and redone the process to obtain render information from a viewport
* Some stats, such as material changes are too difficult to guess on Vulkan, were removed.
* Separated visible and shadow stats, which causes confusion.
* Texture, buffer and general video memory can be queried now.
* Fixed the performance metrics too.
2021-07-03 10:15:04 -03:00
Bastiaan Olij
e86aeb85ce Properly override virtual functions in render_scene_render_rd and subclasses 2021-07-03 20:32:40 +10:00
Rémi Verschelde
3f613236e0
Merge pull request #49576 from clayjohn/VULKAN-SSAO-uniform-sets
Store SSAO uniform sets per viewport
2021-06-14 19:05:19 +02:00
clayjohn
fc93e10092 Store SSAO uniform sets per viewport 2021-06-13 10:42:10 -07:00
Bastiaan Olij
15c1a76361 Add stereoscopic rendering through multiview 2021-06-13 22:52:20 +10:00
reduz
32625145c8 Rename GI Classes
* GIProbe is now VoxelGI
* BakedLightmap is now LightmapGI

As godot adds more ways to provide GI (as an example, SDFGI in 4.0), the different techniques (which have different pros/cons) need to be properly named to avoid confusion.
2021-06-05 09:28:56 -03:00
Aaron Franke
de3f6699a5
Rename Transform to Transform3D in core 2021-06-03 07:30:01 -04:00
Bastiaan Olij
308b26762b Split RenderDataRD struct from RenderState struct to simplify passing our render data around the renderer. 2021-05-10 21:52:15 +10:00
jfons
6995b0429c Assorted fixes to UV unwrapping and GPU lightmapper
Various fixes to UV2 unwrapping and the GPU lightmapper. Listed here for
context in case of git blame/bisect:

* Fix UV2 unwrapping on import, also cleaned up the unwrap cache code.
* Fix saving of RGBA images in EXR format.
* Fixes to the GPU lightmapper:
	- Added padding between atlas elements, avoids bleeding.
	- Remove old SDF generation code.
	- Fix baked attenuation for Omni/Spot lights.
	- Fix baking of material properties onto UV2 (wireframe was
	  wrongly used before).
	- Disable statically baked lights for objects that have a
	  lightmap texture to avoid applying the same light twice.
	- Fix lightmap pairing in RendererSceneCull.
	- Fix UV2 array generated from `RenderingServer::mesh_surface_get_arrays()`.
	- Port autoexposure fix for OIDN from 3.x.
	- Save debug textures as EXR when using floating point format.
2021-05-03 18:10:34 +02:00
Bastiaan Olij
58ff0dac1a Create mobile renderer 2021-05-03 21:54:11 +10:00
jfons
4d9d99bb82 Implement occlusion culling
Added an occlusion culling system with support for static occluder meshes.
It can be enabled via `Project Settings > Rendering > Occlusion Culling > Use Occlusion Culling`.

Occluders are defined via the new `Occluder3D` resource and instanced using the new
`OccluderInstance3D` node. The occluders can also be automatically baked from a
scene using the built-in editor plugin.
2021-04-23 21:45:23 +02:00
Bastiaan Olij
ebf10fe736 Fix reflection probe 2021-04-21 23:42:29 +10:00
Bastiaan Olij
71c5d73b5e Remove low_end option from renderer, being replaced by separate implementation 2021-04-09 22:04:56 +10:00
Bastiaan Olij
dad40fa2df Move clustered renderer functionality 2021-04-05 23:11:38 +10:00
Bastiaan Olij
f7cbeb0d1b Removed time duplicate from ForwardClustered. Just use the variable from its superclass 2021-03-11 18:08:10 +11:00
Bastiaan Olij
dfc41eb6f6 Moving RendererStorageRD *storage to protected, no sense duplicating it in forward_clustered 2021-03-11 13:24:23 +11:00
Bastiaan Olij
760be46a81 Renamed SDGIShader to SDFGIShader and moved a bunch of things to private 2021-03-04 11:53:09 +11:00
Bastiaan Olij
1f96ba5141 Moving GI code into RendererServerGIRD
Moving Skyshader code into RendererServerSkyRD
2021-02-25 15:15:48 +11:00
reduz
8b19ffd810 Make Servers truly Thread Safe
-Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread).
-RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault.
-Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory).
-3D physics server changed to be made multithread friendly.
-Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads.
-Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
2021-02-10 13:21:46 -03:00
reduz
3e2281a347 Improve SDFGI indirect light feedback loop
-Use occlusion for feedback, further reduces light leaking.
-More control on feedback, now its a slider.
2021-02-07 21:30:12 -03:00
jfons
99e1ce0690 Invert spotlight angle attenuation
Inverted the spotlight angle attenuation so a higher value results in
a dimmer light, this makes it more consistent with the distance
attenuation.

Also changed the way spotlighs are computed in SDFGI
and GIPorbes and GPU lightmapper, now it matches the falloff used in the scene rendering
code.
2021-02-07 20:10:33 +01:00
reduz
8faf23b52b Simplify Volumetric Fog
-Always use temporal reproject, it just loos way better than any other filter.
-By always using termporal reproject, the shadowmap reduction can be done away with, massively improving performance.
-Disadvantage of temporal reproject is update latency so..
-Made sure a gaussian filter runs in XY after fog, this allows to keep stability and lower latency.
2021-02-06 15:08:21 -03:00
reduz
7997544af5 Added temporal reprojection to Volumetric Fog
-It's an option, just enable it
-Just works, don't have to do anything else.
2021-02-05 10:52:54 -03:00
reduz
f20999f6fe Rewrote how barriers work for faster rendering
-Added more finegrained control in RenderingDevice API
-Optimized barriers (use less ones for thee same)
-General optimizations
-Shadows render all together unbarriered
-GI can render together with shadows.
-SDFGI can render together with depth-preoass.
-General fixes
-Added GPU detection
2021-02-04 09:42:28 -03:00
reduz
a9beb7aa8c Shadow map rendering optimization
-All shadow rendering is done with raster now (no compute)
-All shadow rendering is done by rendering directly to the shadow atlas
-Improved how buffer clearing is done to optimize the above.
-Ability to set shadows as 16 bits.
2021-01-24 20:17:28 -03:00
reduz
6fe342478b Several GI related optimizations and fixes
-SDFGI direct light is done over many frames
-SDFGI Changed settings for rays/frame
-SDFGI Misc optimizations
-SDFGI Bug fix on probe scroll

-GIProbe was not working, got it to work again
-GIProbe dynamic objects were not working, fixed

-Added a half size GI option.
2021-01-24 19:29:07 -03:00
reduz
099dee35f4 Added GPU based cluster builder
Clustering is now GPU based, uses an implementation based on the Activision algorithm.
2021-01-19 23:31:06 +01:00
reduz
7008e3c6ea Shader optimizations to reduce VGPR usage and increase occupancy 2021-01-19 23:31:02 +01:00
reduz
cdb216f4e4 Added ability to visualize native shaders 2021-01-06 09:40:09 -03:00
reduz
5d2a1d7892 Rewrite render code to be more cache and thread friendly. 2021-01-05 08:50:18 -03:00
reduz
9a2f18f8e7 Reorganize renderer code.
So it can hopefully be made more cache efficient afterwards.
2021-01-05 08:50:18 -03:00
Rémi Verschelde
b5334d14f7
Update copyright statements to 2021
Happy new year to the wonderful Godot community!

2020 has been a tough year for most of us personally, but a good year for
Godot development nonetheless with a huge amount of work done towards Godot
4.0 and great improvements backported to the long-lived 3.2 branch.

We've had close to 400 contributors to engine code this year, authoring near
7,000 commit! (And that's only for the `master` branch and for the engine code,
there's a lot more when counting docs, demos and other first-party repos.)

Here's to a great year 2021 for all Godot users 🎆
2021-01-01 20:19:21 +01:00
reduz
548524152e Rewrite culling to be more cache/thread friendly.
-Uses a single array with all data
-Massive performance improvement
-Does not support threads yet, but code is now thread friendly
2020-12-26 19:11:33 -03:00
reduz
83058597cf Replace Octree by DynamicBVH in cull code
-Much greater pairing/unpairing performance
-For now, using it for culling too, but this will change in a couple of days.
-Added a paged allocator, to efficiently alloc/free some types of objects.
2020-12-23 19:31:30 -03:00
clayjohn
44f8922305 Port ASSAO to Godot to replace SAO 2020-12-21 23:08:59 -08:00
reduz
d2302f53d6 Implement automatic LOD (Level of Detail)
-Happens on import by default for all models
-Just works (tm)
-Biasing can be later adjusted per node or per viewport (as well as globally)
-Disabled AABB.get_support test because its broken
2020-12-18 15:48:03 +01:00
reduz
2748b9a10d Add support for low-end 3D rendering.
-Reduce number of uniform sets from 6 to 4.
-Remove features in low end mode, in order to reduce the number of texture units fit to 16.
2020-12-07 20:50:57 -03:00
reduz
2787ad65be RenderingServer reorganization 2020-12-04 18:39:46 -03:00