godotengine
/
godot
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #39068 from lawnjelly/kessel_settings_rename 

Rename batching project settings in preparation for GLES3
This commit is contained in:
Rémi Verschelde 2020-06-07 14:25:50 +02:00 committed by GitHub
parent 776eb76018 c7d18dec70
commit 6e1af78df4
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 68 additions and 63 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

68
doc/classes/ProjectSettings.xml
View File

@ -970,48 +970,48 @@
Fix to improve physics jitter, specially on monitors where refresh rate is different than the physics FPS.
[b]Note:[/b] This property is only read when the project starts. To change the physics FPS at runtime, set [member Engine.physics_jitter_fix] instead.
</member>
<member name="rendering/batching/debug/diagnose_frame" type="bool" setter="" getter="" default="false">
When batching is on, this regularly prints a frame diagnosis log. Note that this will degrade performance.
</member>
<member name="rendering/batching/debug/flash_batching" type="bool" setter="" getter="" default="false">
[b]Experimental[/b] For regression testing against the old renderer. If this is switched on, and [code]use_batching[/code] is set, the renderer will swap alternately between using the old renderer, and the batched renderer, on each frame. This makes it easy to identify visual differences. Performance will be degraded.
</member>
<member name="rendering/batching/lights/max_join_items" type="int" setter="" getter="" default="32">
Lights have the potential to prevent joining items, and break many of the performance benefits of batching. This setting enables some complex logic to allow joining items if their lighting is similar, and overlap tests pass. This can significantly improve performance in some games. Set to 0 to switch off. With large values the cost of overlap tests may lead to diminishing returns.
</member>
<member name="rendering/batching/lights/scissor_area_threshold" type="float" setter="" getter="" default="1.0">
Sets the proportion of the total screen area (in pixels) that must be saved by a scissor operation in order to activate light scissoring. This can prevent parts of items being rendered outside the light area. Lower values scissor more aggressively. A value of 1 scissors none of the items, a value of 0 scissors every item. The power of 4 of the value is used, in order to emphasize the lower range, and multiplied by the total screen area in pixels to give the threshold. This can reduce fill rate requirements in scenes with a lot of lighting.
</member>
<member name="rendering/batching/options/single_rect_fallback" type="bool" setter="" getter="" default="false">
Enabling this setting uses the legacy method to draw batches containing only one rect. The legacy method is faster (approx twice as fast), but can cause flicker on some systems. In order to directly compare performance with the non-batching renderer you can set this to true, but it is recommended to turn this off unless you can guarantee your target hardware will work with this method.
</member>
<member name="rendering/batching/options/use_batching" type="bool" setter="" getter="" default="true">
Turns batching on and off. Batching increases performance by reducing the amount of graphics API drawcalls.
</member>
<member name="rendering/batching/options/use_batching_in_editor" type="bool" setter="" getter="" default="true">
Switches on batching within the editor.
</member>
<member name="rendering/batching/parameters/batch_buffer_size" type="int" setter="" getter="" default="16384">
Size of buffer reserved for batched vertices. Larger size enables larger batches, but there are diminishing returns for the memory used. This should only have a minor effect on performance.
</member>
<member name="rendering/batching/parameters/colored_vertex_format_threshold" type="float" setter="" getter="" default="0.25">
Including color in the vertex format has a cost, however, not including color prevents batching across color changes. This threshold determines the ratio of [code]number of vertex color changes / total number of vertices[/code] above which vertices will be translated to colored format. A value of 0 will always use colored vertices, 1 will never use colored vertices.
</member>
<member name="rendering/batching/parameters/item_reordering_lookahead" type="int" setter="" getter="" default="4">
In certain circumstances, the batcher can reorder items in order to better join them. This may result in better performance. An overlap test is needed however for each item lookahead, so there is a trade off, with diminishing returns. If you are getting no benefit, setting this to 0 will switch it off.
</member>
<member name="rendering/batching/parameters/max_join_item_commands" type="int" setter="" getter="" default="16">
Sets the number of commands to lookahead to determine whether to batch render items. A value of 1 can join items consisting of single commands, 0 turns off joining. Higher values are in theory more likely to join, however this has diminishing returns and has a runtime cost so a small value is recommended.
</member>
<member name="rendering/environment/default_clear_color" type="Color" setter="" getter="" default="Color( 0.3, 0.3, 0.3, 1 )">
Default background clear color. Overridable per [Viewport] using its [Environment]. See [member Environment.background_mode] and [member Environment.background_color] in particular. To change this default color programmatically, use [method VisualServer.set_default_clear_color].
</member>
<member name="rendering/environment/default_environment" type="String" setter="" getter="" default="&quot;&quot;">
[Environment] that will be used as a fallback environment in case a scene does not specify its own environment. The default environment is loaded in at scene load time regardless of whether you have set an environment or not. If you do not rely on the fallback environment, it is best to delete [code]default_env.tres[/code], or to specify a different default environment here.
</member>
<member name="rendering/gles2/batching/batch_buffer_size" type="int" setter="" getter="" default="16384">
Size of buffer reserved for batched vertices. Larger size enables larger batches, but there are diminishing returns for the memory used.
</member>
<member name="rendering/gles2/batching/colored_vertex_format_threshold" type="float" setter="" getter="" default="0.25">
Including color in the vertex format has a cost, however, not including color prevents batching across color changes. This threshold determines the ratio of [code]number of vertex color changes / total number of vertices[/code] above which vertices will be translated to colored format. A value of 0 will always use colored vertices, 1 will never use colored vertices.
</member>
<member name="rendering/gles2/batching/item_reordering_lookahead" type="int" setter="" getter="" default="4">
In certain circumstances, the batcher can reorder items in order to better join them. This may result in better performance. An overlap test is needed however for each item lookahead, so there is a trade off, with diminishing returns. If you are getting no benefit, setting this to 0 will switch it off.
</member>
<member name="rendering/gles2/batching/light_max_join_items" type="int" setter="" getter="" default="32">
Lights have the potential to prevent joining items, and break many of the performance benefits of batching. This setting enables some complex logic to allow joining items if their lighting is similar, and overlap tests pass. This can significantly improve performance in some games. Set to 0 to switch off. With large values the cost of overlap tests may lead to diminishing returns.
</member>
<member name="rendering/gles2/batching/light_scissor_area_threshold" type="float" setter="" getter="" default="1.0">
Sets the proportion of the screen area that must be saved by a scissor operation in order to activate light scissoring. This can prevent parts of items being rendered outside the light area. Lower values scissor more aggressively. A value of 1 scissors none of the items, a value of 0 scissors every item. This can reduce fill rate requirements in scenes with a lot of lighting.
</member>
<member name="rendering/gles2/batching/max_join_item_commands" type="int" setter="" getter="" default="16">
Sets the number of commands to lookahead to determine whether to batch render items. A value of 1 can join items consisting of single commands, 0 turns off joining. Higher values are in theory more likely to join, however this has diminishing returns and has a runtime cost so a small value is recommended.
</member>
<member name="rendering/gles2/batching/single_rect_fallback" type="bool" setter="" getter="" default="false">
Enabling this uses the legacy method to draw single rects, which is faster, but can cause flicker on some systems. This is best disabled unless crucial for performance.
</member>
<member name="rendering/gles2/batching/use_batching" type="bool" setter="" getter="" default="true">
Turns batching on and off. Batching increases performance by reducing the amount of graphics API drawcalls.
</member>
<member name="rendering/gles2/debug/diagnose_frame" type="bool" setter="" getter="" default="false">
When batching is on, this regularly prints a frame diagnosis log. Note that this will degrade performance.
</member>
<member name="rendering/gles2/debug/disable_half_float" type="bool" setter="" getter="" default="false">
<member name="rendering/gles2/compatibility/disable_half_float" type="bool" setter="" getter="" default="false">
The use of half-float vertex compression may be producing rendering errors on some platforms (especially iOS). These have been seen particularly in particles. Disabling half-float may resolve these problems.
</member>
<member name="rendering/gles2/debug/flash_batching" type="bool" setter="" getter="" default="false">
[b]Experimental[/b] For regression testing against the old renderer. If this is switched on, and [code]use_batching[/code] is set, the renderer will swap alternately between using the old renderer, and the batched renderer, on each frame. This makes it easy to identify visual differences. Performance will be degraded.
</member>
<member name="rendering/gles2/debug/use_batching_in_editor" type="bool" setter="" getter="" default="true">
[b]Experimental[/b] Switches on batching within the editor. Use with caution - note that if your editor does not render correctly you may need to edit your [code]project.godot[/code] and remove the use_batching_in_editor setting manually.
</member>
<member name="rendering/limits/buffers/blend_shape_max_buffer_size_kb" type="int" setter="" getter="" default="4096">
Max buffer size for blend shapes. Any blend shape bigger than this will not work.
</member>

27
drivers/gles2/rasterizer_canvas_gles2.cpp
View File

@ -3339,30 +3339,35 @@ void RasterizerCanvasGLES2::_calculate_scissor_threshold_area() {
void RasterizerCanvasGLES2::initialize() {
RasterizerCanvasBaseGLES2::initialize();
bdata.settings_use_batching = GLOBAL_GET("rendering/gles2/batching/use_batching");
bdata.settings_max_join_item_commands = GLOBAL_GET("rendering/gles2/batching/max_join_item_commands");
bdata.settings_colored_vertex_format_threshold = GLOBAL_GET("rendering/gles2/batching/colored_vertex_format_threshold");
bdata.settings_item_reordering_lookahead = GLOBAL_GET("rendering/gles2/batching/item_reordering_lookahead");
bdata.settings_light_max_join_items = GLOBAL_GET("rendering/gles2/batching/light_max_join_items");
bdata.settings_use_single_rect_fallback = GLOBAL_GET("rendering/gles2/batching/single_rect_fallback");
bdata.settings_use_batching = GLOBAL_GET("rendering/batching/options/use_batching");
bdata.settings_max_join_item_commands = GLOBAL_GET("rendering/batching/parameters/max_join_item_commands");
bdata.settings_colored_vertex_format_threshold = GLOBAL_GET("rendering/batching/parameters/colored_vertex_format_threshold");
bdata.settings_item_reordering_lookahead = GLOBAL_GET("rendering/batching/parameters/item_reordering_lookahead");
bdata.settings_light_max_join_items = GLOBAL_GET("rendering/batching/lights/max_join_items");
bdata.settings_use_single_rect_fallback = GLOBAL_GET("rendering/batching/options/single_rect_fallback");
// we can use the threshold to determine whether to turn scissoring off or on
bdata.settings_scissor_threshold = GLOBAL_GET("rendering/gles2/batching/light_scissor_area_threshold");
bdata.settings_scissor_threshold = GLOBAL_GET("rendering/batching/lights/scissor_area_threshold");
if (bdata.settings_scissor_threshold > 0.999f) {
bdata.settings_scissor_lights = false;
} else {
bdata.settings_scissor_lights = true;
// apply power of 4 relationship for the area, as most of the important changes
// will be happening at low values of scissor threshold
bdata.settings_scissor_threshold *= bdata.settings_scissor_threshold;
bdata.settings_scissor_threshold *= bdata.settings_scissor_threshold;
}
// The sweet spot on my desktop for cache is actually smaller than the max, and this
// is the default. This saves memory too so we will use it for now, needs testing to see whether this varies according
// to device / platform.
bdata.settings_batch_buffer_num_verts = GLOBAL_GET("rendering/gles2/batching/batch_buffer_size");
bdata.settings_batch_buffer_num_verts = GLOBAL_GET("rendering/batching/parameters/batch_buffer_size");
// override the use_batching setting in the editor
// (note that if the editor can't start, you can't change the use_batching project setting!)
if (Engine::get_singleton()->is_editor_hint()) {
bool use_in_editor = GLOBAL_GET("rendering/gles2/debug/use_batching_in_editor");
bool use_in_editor = GLOBAL_GET("rendering/batching/options/use_batching_in_editor");
bdata.settings_use_batching = use_in_editor;
// fix some settings in the editor, as the performance not worth the risk
@ -3383,7 +3388,7 @@ void RasterizerCanvasGLES2::initialize() {
// This should not be used except during development.
// make a note of the original choice in case we are flashing on and off the batching
bdata.settings_use_batching_original_choice = bdata.settings_use_batching;
bdata.settings_flash_batching = GLOBAL_GET("rendering/gles2/debug/flash_batching");
bdata.settings_flash_batching = GLOBAL_GET("rendering/batching/debug/flash_batching");
if (!bdata.settings_use_batching) {
// no flash when batching turned off
bdata.settings_flash_batching = false;
@ -3392,7 +3397,7 @@ void RasterizerCanvasGLES2::initialize() {
// frame diagnosis. print out the batches every nth frame
bdata.settings_diagnose_frame = false;
if (!Engine::get_singleton()->is_editor_hint() && bdata.settings_use_batching) {
bdata.settings_diagnose_frame = GLOBAL_GET("rendering/gles2/debug/diagnose_frame");
bdata.settings_diagnose_frame = GLOBAL_GET("rendering/batching/debug/diagnose_frame");
}
// the maximum num quads in a batch is limited by GLES2. We can have only 16 bit indices,

36
servers/visual_server.cpp
View File

@ -2420,25 +2420,25 @@ VisualServer::VisualServer() {
GLOBAL_DEF("rendering/quality/filters/use_nearest_mipmap_filter", false);
GLOBAL_DEF("rendering/gles2/batching/use_batching", true);
GLOBAL_DEF("rendering/gles2/batching/max_join_item_commands", 16);
GLOBAL_DEF("rendering/gles2/batching/colored_vertex_format_threshold", 0.25f);
GLOBAL_DEF("rendering/gles2/batching/light_scissor_area_threshold", 1.0f);
GLOBAL_DEF("rendering/gles2/batching/light_max_join_items", 32);
GLOBAL_DEF("rendering/gles2/batching/batch_buffer_size", 16384);
GLOBAL_DEF("rendering/gles2/batching/item_reordering_lookahead", 4);
GLOBAL_DEF("rendering/gles2/batching/single_rect_fallback", false);
GLOBAL_DEF("rendering/gles2/debug/flash_batching", false);
GLOBAL_DEF("rendering/gles2/debug/diagnose_frame", false);
GLOBAL_DEF_RST("rendering/gles2/debug/use_batching_in_editor", true);
GLOBAL_DEF("rendering/gles2/debug/disable_half_float", false);
GLOBAL_DEF("rendering/batching/options/use_batching", true);
GLOBAL_DEF_RST("rendering/batching/options/use_batching_in_editor", true);
GLOBAL_DEF("rendering/batching/options/single_rect_fallback", false);
GLOBAL_DEF("rendering/batching/parameters/max_join_item_commands", 16);
GLOBAL_DEF("rendering/batching/parameters/colored_vertex_format_threshold", 0.25f);
GLOBAL_DEF("rendering/batching/lights/scissor_area_threshold", 1.0f);
GLOBAL_DEF("rendering/batching/lights/max_join_items", 32);
GLOBAL_DEF("rendering/batching/parameters/batch_buffer_size", 16384);
GLOBAL_DEF("rendering/batching/parameters/item_reordering_lookahead", 4);
GLOBAL_DEF("rendering/batching/debug/flash_batching", false);
GLOBAL_DEF("rendering/batching/debug/diagnose_frame", false);
GLOBAL_DEF("rendering/gles2/compatibility/disable_half_float", false);
ProjectSettings::get_singleton()->set_custom_property_info("rendering/gles2/batching/max_join_item_commands", PropertyInfo(Variant::INT, "rendering/gles2/batching/max_join_item_commands", PROPERTY_HINT_RANGE, "0,65535"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/gles2/batching/colored_vertex_format_threshold", PropertyInfo(Variant::REAL, "rendering/gles2/batching/colored_vertex_format_threshold", PROPERTY_HINT_RANGE, "0.0,1.0,0.01"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/gles2/batching/batch_buffer_size", PropertyInfo(Variant::INT, "rendering/gles2/batching/batch_buffer_size", PROPERTY_HINT_RANGE, "1024,65535,1024"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/gles2/batching/light_scissor_area_threshold", PropertyInfo(Variant::REAL, "rendering/gles2/batching/light_scissor_area_threshold", PROPERTY_HINT_RANGE, "0.0,1.0"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/gles2/batching/light_max_join_items", PropertyInfo(Variant::INT, "rendering/gles2/batching/light_max_join_items", PROPERTY_HINT_RANGE, "0,512"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/gles2/batching/item_reordering_lookahead", PropertyInfo(Variant::INT, "rendering/gles2/batching/item_reordering_lookahead", PROPERTY_HINT_RANGE, "0,256"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/batching/parameters/max_join_item_commands", PropertyInfo(Variant::INT, "rendering/batching/parameters/max_join_item_commands", PROPERTY_HINT_RANGE, "0,65535"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/batching/parameters/colored_vertex_format_threshold", PropertyInfo(Variant::REAL, "rendering/batching/parameters/colored_vertex_format_threshold", PROPERTY_HINT_RANGE, "0.0,1.0,0.01"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/batching/parameters/batch_buffer_size", PropertyInfo(Variant::INT, "rendering/batching/parameters/batch_buffer_size", PROPERTY_HINT_RANGE, "1024,65535,1024"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/batching/lights/scissor_area_threshold", PropertyInfo(Variant::REAL, "rendering/batching/lights/scissor_area_threshold", PROPERTY_HINT_RANGE, "0.0,1.0"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/batching/lights/max_join_items", PropertyInfo(Variant::INT, "rendering/batching/lights/max_join_items", PROPERTY_HINT_RANGE, "0,512"));
ProjectSettings::get_singleton()->set_custom_property_info("rendering/batching/parameters/item_reordering_lookahead", PropertyInfo(Variant::INT, "rendering/batching/parameters/item_reordering_lookahead", PROPERTY_HINT_RANGE, "0,256"));
}
VisualServer::~VisualServer() {