asgardius/virtualx-engine
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
virtualx-engine/servers/rendering/renderer_storage.h

415 lines
20 KiB
C++
Raw Normal View History

RenderingServer reorganization
2020-12-04 19:26:24 +01:00
/*************************************************************************/
/* renderer_storage.h */
/*************************************************************************/
/* This file is part of: */
/* GODOT ENGINE */
/* https://godotengine.org */
/*************************************************************************/
Update copyright statements to 2022 Happy new year to the wonderful Godot community!
2022-01-03 21:27:34 +01:00
/* Copyright (c) 2007-2022 Juan Linietsky, Ariel Manzur. */
/* Copyright (c) 2014-2022 Godot Engine contributors (cf. AUTHORS.md). */
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
/* */
/* Permission is hereby granted, free of charge, to any person obtaining */
/* a copy of this software and associated documentation files (the */
/* "Software"), to deal in the Software without restriction, including */
/* without limitation the rights to use, copy, modify, merge, publish, */
/* distribute, sublicense, and/or sell copies of the Software, and to */
/* permit persons to whom the Software is furnished to do so, subject to */
/* the following conditions: */
/* */
/* The above copyright notice and this permission notice shall be */
/* included in all copies or substantial portions of the Software. */
/* */
/* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, */
/* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF */
/* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.*/
/* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY */
/* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, */
/* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE */
/* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. */
/*************************************************************************/
#ifndef RENDERINGSERVERSTORAGE_H
#define RENDERINGSERVERSTORAGE_H
#include "servers/rendering_server.h"
class RendererStorage {
Color default_clear_color;
public:
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
enum DependencyChangedNotification {
DEPENDENCY_CHANGED_AABB,
DEPENDENCY_CHANGED_MATERIAL,
DEPENDENCY_CHANGED_MESH,
DEPENDENCY_CHANGED_MULTIMESH,
DEPENDENCY_CHANGED_MULTIMESH_VISIBLE_INSTANCES,
Implement Particle Trails -Enable the trails and set the length in seconds -Provide a mesh with a skeleton and a skin -Or, alternatively use one of the built-in TubeTrailMesh/RibbonTrailMesh -Works deterministically -Fixed particle collisions (were broken) -Not working in 2D yet (that will happen next)
2021-04-27 17:43:49 +02:00
DEPENDENCY_CHANGED_PARTICLES,
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
DEPENDENCY_CHANGED_DECAL,
DEPENDENCY_CHANGED_SKELETON_DATA,
DEPENDENCY_CHANGED_SKELETON_BONES,
DEPENDENCY_CHANGED_LIGHT,
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.
2021-07-13 01:32:05 +02:00
DEPENDENCY_CHANGED_LIGHT_SOFT_SHADOW_AND_PROJECTOR,
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
DEPENDENCY_CHANGED_REFLECTION_PROBE,
};
struct DependencyTracker;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
struct Dependency {
void changed_notify(DependencyChangedNotification p_notification);
void deleted_notify(const RID &p_rid);
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
~Dependency();
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
private:
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
friend struct DependencyTracker;
Map<DependencyTracker *, uint32_t> instances;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
};
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
struct DependencyTracker {
void *userdata = nullptr;
typedef void (*ChangedCallback)(DependencyChangedNotification, DependencyTracker *);
typedef void (*DeletedCallback)(const RID &, DependencyTracker *);
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
ChangedCallback changed_callback = nullptr;
DeletedCallback deleted_callback = nullptr;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
void update_begin() { // call before updating dependencies
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
instance_version++;
}
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
void update_dependency(Dependency *p_dependency) { //called internally, can't be used directly, use update functions in Storage
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
dependencies.insert(p_dependency);
p_dependency->instances[this] = instance_version;
}
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
void update_end() { //call after updating dependencies
List<Pair<Dependency *, Map<DependencyTracker *, uint32_t>::Element *>> to_clean_up;
for (Set<Dependency *>::Element *E = dependencies.front(); E; E = E->next()) {
Dependency *dep = E->get();
Map<DependencyTracker *, uint32_t>::Element *F = dep->instances.find(this);
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
ERR_CONTINUE(!F);
if (F->get() != instance_version) {
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
Pair<Dependency *, Map<DependencyTracker *, uint32_t>::Element *> p;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
p.first = dep;
p.second = F;
to_clean_up.push_back(p);
}
}
while (to_clean_up.size()) {
to_clean_up.front()->get().first->instances.erase(to_clean_up.front()->get().second);
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
2021-02-02 20:51:36 +01:00
dependencies.erase(to_clean_up.front()->get().first);
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
to_clean_up.pop_front();
}
}
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
void clear() { // clear all dependencies
for (Set<Dependency *>::Element *E = dependencies.front(); E; E = E->next()) {
Dependency *dep = E->get();
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
dep->instances.erase(this);
}
dependencies.clear();
}
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
~DependencyTracker() { clear(); }
private:
Threaded optimizations to cull and render -Reorganize thread work pool for rendering -Fixes to make secondary command buffers to work (disabled because they need more testing)
2021-01-04 21:00:44 +01:00
friend struct Dependency;
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
uint32_t instance_version = 0;
Set<Dependency *> dependencies;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
};
/* Light API */
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).
2021-02-09 17:19:03 +01:00
virtual RID directional_light_allocate() = 0;
virtual void directional_light_initialize(RID p_rid) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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).
2021-02-09 17:19:03 +01:00
virtual RID omni_light_allocate() = 0;
virtual void omni_light_initialize(RID p_rid) = 0;
virtual RID spot_light_allocate() = 0;
virtual void spot_light_initialize(RID p_rid) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void light_set_color(RID p_light, const Color &p_color) = 0;
virtual void light_set_param(RID p_light, RS::LightParam p_param, float p_value) = 0;
virtual void light_set_shadow(RID p_light, bool p_enabled) = 0;
virtual void light_set_projector(RID p_light, RID p_texture) = 0;
virtual void light_set_negative(RID p_light, bool p_enable) = 0;
virtual void light_set_cull_mask(RID p_light, uint32_t p_mask) = 0;
Implement distance fade properties in OmniLight3D and SpotLight3D This can be used to fade lights and their shadows in the distance, similar to Decal nodes. This can bring significant performance improvements, especially for lights with shadows enabled and when using higher-than-default shadow quality settings. While lights can be smoothly faded out over distance, shadows are currently "all or nothing" since per-light shadow color is no longer customizable in the Vulkan renderer. This may result in noticeable pop-in when leaving the shadow cutoff distance, but depending on the scene, it may not always be that noticeable.
2022-02-24 22:55:14 +01:00
virtual void light_set_distance_fade(RID p_light, bool p_enabled, float p_begin, float p_shadow, float p_length) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void light_set_reverse_cull_face_mode(RID p_light, bool p_enabled) = 0;
virtual void light_set_bake_mode(RID p_light, RS::LightBakeMode p_bake_mode) = 0;
virtual void light_set_max_sdfgi_cascade(RID p_light, uint32_t p_cascade) = 0;
virtual void light_omni_set_shadow_mode(RID p_light, RS::LightOmniShadowMode p_mode) = 0;
virtual void light_directional_set_shadow_mode(RID p_light, RS::LightDirectionalShadowMode p_mode) = 0;
virtual void light_directional_set_blend_splits(RID p_light, bool p_enable) = 0;
virtual bool light_directional_get_blend_splits(RID p_light) const = 0;
Replace DirectionalLight3D's `use_in_sky_only` with `sky_mode` enum 3 options are available: - Light and Sky (default) - Light Only (new) - Sky Only (equivalent to `use_in_sky_only = true`) Co-authored by: clayjohn <claynjohn@gmail.com>
2021-10-20 00:40:46 +02:00
virtual void light_directional_set_sky_mode(RID p_light, RS::LightDirectionalSkyMode p_mode) = 0;
virtual RS::LightDirectionalSkyMode light_directional_get_sky_mode(RID p_light) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual RS::LightDirectionalShadowMode light_directional_get_shadow_mode(RID p_light) = 0;
virtual RS::LightOmniShadowMode light_omni_get_shadow_mode(RID p_light) = 0;
virtual bool light_has_shadow(RID p_light) const = 0;
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.
2021-07-13 01:32:05 +02:00
virtual bool light_has_projector(RID p_light) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual RS::LightType light_get_type(RID p_light) const = 0;
virtual AABB light_get_aabb(RID p_light) const = 0;
virtual float light_get_param(RID p_light, RS::LightParam p_param) = 0;
virtual Color light_get_color(RID p_light) = 0;
virtual RS::LightBakeMode light_get_bake_mode(RID p_light) = 0;
virtual uint32_t light_get_max_sdfgi_cascade(RID p_light) = 0;
virtual uint64_t light_get_version(RID p_light) const = 0;
/* PROBE API */
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).
2021-02-09 17:19:03 +01:00
virtual RID reflection_probe_allocate() = 0;
virtual void reflection_probe_initialize(RID p_rid) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void reflection_probe_set_update_mode(RID p_probe, RS::ReflectionProbeUpdateMode p_mode) = 0;
virtual void reflection_probe_set_resolution(RID p_probe, int p_resolution) = 0;
virtual void reflection_probe_set_intensity(RID p_probe, float p_intensity) = 0;
virtual void reflection_probe_set_ambient_mode(RID p_probe, RS::ReflectionProbeAmbientMode p_mode) = 0;
virtual void reflection_probe_set_ambient_color(RID p_probe, const Color &p_color) = 0;
virtual void reflection_probe_set_ambient_energy(RID p_probe, float p_energy) = 0;
virtual void reflection_probe_set_max_distance(RID p_probe, float p_distance) = 0;
virtual void reflection_probe_set_extents(RID p_probe, const Vector3 &p_extents) = 0;
virtual void reflection_probe_set_origin_offset(RID p_probe, const Vector3 &p_offset) = 0;
virtual void reflection_probe_set_as_interior(RID p_probe, bool p_enable) = 0;
virtual void reflection_probe_set_enable_box_projection(RID p_probe, bool p_enable) = 0;
virtual void reflection_probe_set_enable_shadows(RID p_probe, bool p_enable) = 0;
virtual void reflection_probe_set_cull_mask(RID p_probe, uint32_t p_layers) = 0;
Rename Lod Threshold to Mesh Lod Threshold This makes it more obvious that the setting only affects mesh LOD, not manual (H)LOD achieved using visibility ranges.
2021-12-29 00:10:41 +01:00
virtual void reflection_probe_set_mesh_lod_threshold(RID p_probe, float p_ratio) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual AABB reflection_probe_get_aabb(RID p_probe) const = 0;
virtual RS::ReflectionProbeUpdateMode reflection_probe_get_update_mode(RID p_probe) const = 0;
virtual uint32_t reflection_probe_get_cull_mask(RID p_probe) const = 0;
virtual Vector3 reflection_probe_get_extents(RID p_probe) const = 0;
virtual Vector3 reflection_probe_get_origin_offset(RID p_probe) const = 0;
virtual float reflection_probe_get_origin_max_distance(RID p_probe) const = 0;
virtual bool reflection_probe_renders_shadows(RID p_probe) const = 0;
Rename Lod Threshold to Mesh Lod Threshold This makes it more obvious that the setting only affects mesh LOD, not manual (H)LOD achieved using visibility ranges.
2021-12-29 00:10:41 +01:00
virtual float reflection_probe_get_mesh_lod_threshold(RID p_probe) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Rewrite render code to be more cache and thread friendly.
2021-01-04 13:33:25 +01:00
virtual void base_update_dependency(RID p_base, DependencyTracker *p_instance) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
/* VOXEL GI API */
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual RID voxel_gi_allocate() = 0;
virtual void voxel_gi_initialize(RID p_rid) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_allocate_data(RID p_voxel_gi, const Transform3D &p_to_cell_xform, const AABB &p_aabb, const Vector3i &p_octree_size, const Vector<uint8_t> &p_octree_cells, const Vector<uint8_t> &p_data_cells, const Vector<uint8_t> &p_distance_field, const Vector<int> &p_level_counts) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual AABB voxel_gi_get_bounds(RID p_voxel_gi) const = 0;
virtual Vector3i voxel_gi_get_octree_size(RID p_voxel_gi) const = 0;
virtual Vector<uint8_t> voxel_gi_get_octree_cells(RID p_voxel_gi) const = 0;
virtual Vector<uint8_t> voxel_gi_get_data_cells(RID p_voxel_gi) const = 0;
virtual Vector<uint8_t> voxel_gi_get_distance_field(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual Vector<int> voxel_gi_get_level_counts(RID p_voxel_gi) const = 0;
virtual Transform3D voxel_gi_get_to_cell_xform(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_dynamic_range(RID p_voxel_gi, float p_range) = 0;
virtual float voxel_gi_get_dynamic_range(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_propagation(RID p_voxel_gi, float p_range) = 0;
virtual float voxel_gi_get_propagation(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_energy(RID p_voxel_gi, float p_energy) = 0;
virtual float voxel_gi_get_energy(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_bias(RID p_voxel_gi, float p_bias) = 0;
virtual float voxel_gi_get_bias(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_normal_bias(RID p_voxel_gi, float p_range) = 0;
virtual float voxel_gi_get_normal_bias(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_interior(RID p_voxel_gi, bool p_enable) = 0;
virtual bool voxel_gi_is_interior(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_use_two_bounces(RID p_voxel_gi, bool p_enable) = 0;
virtual bool voxel_gi_is_using_two_bounces(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual void voxel_gi_set_anisotropy_strength(RID p_voxel_gi, float p_strength) = 0;
virtual float voxel_gi_get_anisotropy_strength(RID p_voxel_gi) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 00:47:26 +02:00
virtual uint32_t voxel_gi_get_version(RID p_probe) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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).
2021-02-09 17:19:03 +01:00
/* LIGHTMAP */
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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).
2021-02-09 17:19:03 +01:00
virtual RID lightmap_allocate() = 0;
virtual void lightmap_initialize(RID p_rid) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void lightmap_set_textures(RID p_lightmap, RID p_light, bool p_uses_spherical_haromics) = 0;
virtual void lightmap_set_probe_bounds(RID p_lightmap, const AABB &p_bounds) = 0;
virtual void lightmap_set_probe_interior(RID p_lightmap, bool p_interior) = 0;
virtual void lightmap_set_probe_capture_data(RID p_lightmap, const PackedVector3Array &p_points, const PackedColorArray &p_point_sh, const PackedInt32Array &p_tetrahedra, const PackedInt32Array &p_bsp_tree) = 0;
virtual PackedVector3Array lightmap_get_probe_capture_points(RID p_lightmap) const = 0;
virtual PackedColorArray lightmap_get_probe_capture_sh(RID p_lightmap) const = 0;
virtual PackedInt32Array lightmap_get_probe_capture_tetrahedra(RID p_lightmap) const = 0;
virtual PackedInt32Array lightmap_get_probe_capture_bsp_tree(RID p_lightmap) const = 0;
virtual AABB lightmap_get_aabb(RID p_lightmap) const = 0;
virtual void lightmap_tap_sh_light(RID p_lightmap, const Vector3 &p_point, Color *r_sh) = 0;
virtual bool lightmap_is_interior(RID p_lightmap) const = 0;
virtual void lightmap_set_probe_capture_update_speed(float p_speed) = 0;
virtual float lightmap_get_probe_capture_update_speed() const = 0;
/* PARTICLES */
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).
2021-02-09 17:19:03 +01:00
virtual RID particles_allocate() = 0;
virtual void particles_initialize(RID p_rid) = 0;
Fixes missng 2D engine bits -Mesh2D now works -MultiMesh2D now works -Polygon2D now works -Added hooks for processing 2D particles -Skeleton2D now works 2D particles still not working, but stuff needed for it is now implemented.
2021-05-10 18:12:44 +02:00
virtual void particles_set_mode(RID p_particles, RS::ParticlesMode p_mode) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_emitting(RID p_particles, bool p_emitting) = 0;
virtual bool particles_get_emitting(RID p_particles) = 0;
virtual void particles_set_amount(RID p_particles, int p_amount) = 0;
Use doubles for time in many other places
2021-02-02 03:16:37 +01:00
virtual void particles_set_lifetime(RID p_particles, double p_lifetime) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_one_shot(RID p_particles, bool p_one_shot) = 0;
Use doubles for time in many other places
2021-02-02 03:16:37 +01:00
virtual void particles_set_pre_process_time(RID p_particles, double p_time) = 0;
Use real_t and double where appropriate in Particles
2021-08-12 16:07:47 +02:00
virtual void particles_set_explosiveness_ratio(RID p_particles, real_t p_ratio) = 0;
virtual void particles_set_randomness_ratio(RID p_particles, real_t p_ratio) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_custom_aabb(RID p_particles, const AABB &p_aabb) = 0;
Use doubles for time in many other places
2021-02-02 03:16:37 +01:00
virtual void particles_set_speed_scale(RID p_particles, double p_scale) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_use_local_coordinates(RID p_particles, bool p_enable) = 0;
virtual void particles_set_process_material(RID p_particles, RID p_material) = 0;
Add Particle Shader Userdata * Adds optional vec4 USERDATA1 .. USERDATA6 to particles, allowing to store custom data. * This data is allocated on demand, so shaders that do not use it do not cost more.
2022-02-14 13:27:10 +01:00
virtual RID particles_get_process_material(RID p_particles) const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_fixed_fps(RID p_particles, int p_fps) = 0;
Implement Particle Trails -Enable the trails and set the length in seconds -Provide a mesh with a skeleton and a skin -Or, alternatively use one of the built-in TubeTrailMesh/RibbonTrailMesh -Works deterministically -Fixed particle collisions (were broken) -Not working in 2D yet (that will happen next)
2021-04-27 17:43:49 +02:00
virtual void particles_set_interpolate(RID p_particles, bool p_enable) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_fractional_delta(RID p_particles, bool p_enable) = 0;
Use real_t and double where appropriate in Particles
2021-08-12 16:07:47 +02:00
virtual void particles_set_collision_base_size(RID p_particles, real_t p_size) = 0;
Implement Particle Trails -Enable the trails and set the length in seconds -Provide a mesh with a skeleton and a skin -Or, alternatively use one of the built-in TubeTrailMesh/RibbonTrailMesh -Works deterministically -Fixed particle collisions (were broken) -Not working in 2D yet (that will happen next)
2021-04-27 17:43:49 +02:00
virtual void particles_set_transform_align(RID p_particles, RS::ParticlesTransformAlign p_transform_align) = 0;
Use real_t and double where appropriate in Particles
2021-08-12 16:07:47 +02:00
virtual void particles_set_trails(RID p_particles, bool p_enable, double p_length) = 0;
Rename Transform to Transform3D in core
2020-10-17 07:08:21 +02:00
virtual void particles_set_trail_bind_poses(RID p_particles, const Vector<Transform3D> &p_bind_poses) = 0;
Implement Particle Trails -Enable the trails and set the length in seconds -Provide a mesh with a skeleton and a skin -Or, alternatively use one of the built-in TubeTrailMesh/RibbonTrailMesh -Works deterministically -Fixed particle collisions (were broken) -Not working in 2D yet (that will happen next)
2021-04-27 17:43:49 +02:00
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_restart(RID p_particles) = 0;
Rename Transform to Transform3D in core
2020-10-17 07:08:21 +02:00
virtual void particles_emit(RID p_particles, const Transform3D &p_transform, const Vector3 &p_velocity, const Color &p_color, const Color &p_custom, uint32_t p_emit_flags) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_set_subemitter(RID p_particles, RID p_subemitter_particles) = 0;
virtual bool particles_is_inactive(RID p_particles) const = 0;
virtual void particles_set_draw_order(RID p_particles, RS::ParticlesDrawOrder p_order) = 0;
virtual void particles_set_draw_passes(RID p_particles, int p_count) = 0;
virtual void particles_set_draw_pass_mesh(RID p_particles, int p_pass, RID p_mesh) = 0;
virtual void particles_request_process(RID p_particles) = 0;
virtual AABB particles_get_current_aabb(RID p_particles) = 0;
virtual AABB particles_get_aabb(RID p_particles) const = 0;
Rename Transform to Transform3D in core
2020-10-17 07:08:21 +02:00
virtual void particles_set_emission_transform(RID p_particles, const Transform3D &p_transform) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual int particles_get_draw_passes(RID p_particles) const = 0;
virtual RID particles_get_draw_pass_mesh(RID p_particles, int p_pass) const = 0;
Implement Particle Trails -Enable the trails and set the length in seconds -Provide a mesh with a skeleton and a skin -Or, alternatively use one of the built-in TubeTrailMesh/RibbonTrailMesh -Works deterministically -Fixed particle collisions (were broken) -Not working in 2D yet (that will happen next)
2021-04-27 17:43:49 +02:00
virtual void particles_set_view_axis(RID p_particles, const Vector3 &p_axis, const Vector3 &p_up_axis) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Reorganize renderer code. So it can hopefully be made more cache efficient afterwards.
2020-12-31 13:42:56 +01:00
virtual void particles_add_collision(RID p_particles, RID p_particles_collision_instance) = 0;
virtual void particles_remove_collision(RID p_particles, RID p_particles_collision_instance) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
Support for 2D particles to collide against SDF -Added SDF collision support for 2D particles -Changed the SDF generation to be fully signed
2021-05-20 16:25:06 +02:00
virtual void particles_set_canvas_sdf_collision(RID p_particles, bool p_enable, const Transform2D &p_xform, const Rect2 &p_to_screen, RID p_texture) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void update_particles() = 0;
/* PARTICLES COLLISION */
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).
2021-02-09 17:19:03 +01:00
virtual RID particles_collision_allocate() = 0;
virtual void particles_collision_initialize(RID p_rid) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_collision_set_collision_type(RID p_particles_collision, RS::ParticlesCollisionType p_type) = 0;
virtual void particles_collision_set_cull_mask(RID p_particles_collision, uint32_t p_cull_mask) = 0;
Use real_t and double where appropriate in Particles
2021-08-12 16:07:47 +02:00
virtual void particles_collision_set_sphere_radius(RID p_particles_collision, real_t p_radius) = 0; //for spheres
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_collision_set_box_extents(RID p_particles_collision, const Vector3 &p_extents) = 0; //for non-spheres
Use real_t and double where appropriate in Particles
2021-08-12 16:07:47 +02:00
virtual void particles_collision_set_attractor_strength(RID p_particles_collision, real_t p_strength) = 0;
virtual void particles_collision_set_attractor_directionality(RID p_particles_collision, real_t p_directionality) = 0;
virtual void particles_collision_set_attractor_attenuation(RID p_particles_collision, real_t p_curve) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual void particles_collision_set_field_texture(RID p_particles_collision, RID p_texture) = 0; //for SDF and vector field, heightfield is dynamic
virtual void particles_collision_height_field_update(RID p_particles_collision) = 0; //for SDF and vector field
virtual void particles_collision_set_height_field_resolution(RID p_particles_collision, RS::ParticlesCollisionHeightfieldResolution p_resolution) = 0; //for SDF and vector field
virtual AABB particles_collision_get_aabb(RID p_particles_collision) const = 0;
virtual bool particles_collision_is_heightfield(RID p_particles_collision) const = 0;
virtual RID particles_collision_get_heightfield_framebuffer(RID p_particles_collision) const = 0;
Addition of FogVolumes, FogShaders, FogMaterial, and overhaul of VolumetricFog Co-authored-by: Brian Semrau <brian.semrau@gmail.com>
2021-10-03 13:28:55 +02:00
/* FOG VOLUMES */
virtual RID fog_volume_allocate() = 0;
virtual void fog_volume_initialize(RID p_rid) = 0;
virtual void fog_volume_set_shape(RID p_fog_volume, RS::FogVolumeShape p_shape) = 0;
virtual void fog_volume_set_extents(RID p_fog_volume, const Vector3 &p_extents) = 0;
virtual void fog_volume_set_material(RID p_fog_volume, RID p_material) = 0;
virtual AABB fog_volume_get_aabb(RID p_fog_volume) const = 0;
virtual RS::FogVolumeShape fog_volume_get_shape(RID p_fog_volume) const = 0;
/* VISIBILITY NOTIFIER */
Refactor VisibilityNotifier3D * This is the 3D counterpart to #49632 * Implemented a bit different as 3D works using instancing After merged, both 2D and 3D classes will most likely be renamed in a separate PR to DisplayNotifier2D/3D.
2021-06-16 20:43:02 +02:00
virtual RID visibility_notifier_allocate() = 0;
virtual void visibility_notifier_initialize(RID p_notifier) = 0;
virtual void visibility_notifier_set_aabb(RID p_notifier, const AABB &p_aabb) = 0;
virtual void visibility_notifier_set_callbacks(RID p_notifier, const Callable &p_enter_callbable, const Callable &p_exit_callable) = 0;
virtual AABB visibility_notifier_get_aabb(RID p_notifier) const = 0;
virtual void visibility_notifier_call(RID p_notifier, bool p_enter, bool p_deferred) = 0;
Reorganize renderer code. So it can hopefully be made more cache efficient afterwards.
2020-12-31 13:42:56 +01:00
//used from 2D and 3D
virtual RID particles_collision_instance_create(RID p_collision) = 0;
Rename Transform to Transform3D in core
2020-10-17 07:08:21 +02:00
virtual void particles_collision_instance_set_transform(RID p_collision_instance, const Transform3D &p_transform) = 0;
Reorganize renderer code. So it can hopefully be made more cache efficient afterwards.
2020-12-31 13:42:56 +01:00
virtual void particles_collision_instance_set_active(RID p_collision_instance, bool p_active) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual RS::InstanceType get_base_type(RID p_rid) const = 0;
virtual bool free(RID p_rid) = 0;
virtual bool has_os_feature(const String &p_feature) const = 0;
virtual void update_dirty_resources() = 0;
virtual void set_debug_generate_wireframes(bool p_generate) = 0;
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 01:14:19 +02:00
virtual void update_memory_info() = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
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 01:14:19 +02:00
virtual uint64_t get_rendering_info(RS::RenderingInfo p_info) = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
virtual String get_video_adapter_name() const = 0;
virtual String get_video_adapter_vendor() const = 0;
Add `RenderingServer.get_video_adapter_type()` method This can be used to distinguish between integrated, dedicated, virtual and software-emulated GPUs. This in turn can be used to automatically adjust graphics settings, or warn users about features that may run slowly on their hardware.
2021-12-10 17:01:51 +01:00
virtual RenderingDevice::DeviceType get_video_adapter_type() const = 0;
RenderingServer reorganization
2020-12-04 19:26:24 +01:00
static RendererStorage *base_singleton;
void set_default_clear_color(const Color &p_color) {
default_clear_color = p_color;
}
Color get_default_clear_color() const {
return default_clear_color;
}
#define TIMESTAMP_BEGIN() \
{ \
if (RSG::storage->capturing_timestamps) \
RSG::storage->capture_timestamps_begin(); \
}
#define RENDER_TIMESTAMP(m_text) \
{ \
if (RSG::storage->capturing_timestamps) \
RSG::storage->capture_timestamp(m_text); \
}
bool capturing_timestamps = false;
virtual void capture_timestamps_begin() = 0;
virtual void capture_timestamp(const String &p_name) = 0;
virtual uint32_t get_captured_timestamps_count() const = 0;
virtual uint64_t get_captured_timestamps_frame() const = 0;
virtual uint64_t get_captured_timestamp_gpu_time(uint32_t p_index) const = 0;
virtual uint64_t get_captured_timestamp_cpu_time(uint32_t p_index) const = 0;
virtual String get_captured_timestamp_name(uint32_t p_index) const = 0;
RendererStorage();
virtual ~RendererStorage() {}
};
#endif // RENDERINGSERVERSTORAGE_H
Reference in a new issue Copy permalink