The work was performed by collaboration of TheForge and Google. I am
merely splitting it up into smaller PRs and cleaning it up.
This is the most "risky" PR so far because the previous ones have been
miscellaneous stuff aimed at either [improve
debugging](https://github.com/godotengine/godot/pull/90993) (e.g. device
lost), [improve Android
experience](https://github.com/godotengine/godot/pull/96439) (add Swappy
for better Frame Pacing + Pre-Transformed Swapchains for slightly better
performance), or harmless [ASTC
improvements](https://github.com/godotengine/godot/pull/96045) (better
performance by simply toggling a feature when available).
However this PR contains larger modifications aimed at improving
performance or reducing memory fragmentation. With greater
modifications, come greater risks of bugs or breakage.
Changes introduced by this PR:
TBDR GPUs (e.g. most of Android + iOS + M1 Apple) support rendering to
Render Targets that are not backed by actual GPU memory (everything
stays in cache). This works as long as load action isn't `LOAD`, and
store action must be `DONT_CARE`. This saves VRAM (it also makes
painfully obvious when a mistake introduces a performance regression).
Of particular usefulness is when doing MSAA and keeping the raw MSAA
content is not necessary.
Some GPUs get faster when the sampler settings are hard-coded into the
GLSL shaders (instead of being dynamically bound at runtime). This
required changes to the GLSL shaders, PSO creation routines, Descriptor
creation routines, and Descriptor binding routines.
- `bool immutable_samplers_enabled = true`
Setting it to false enforces the old behavior. Useful for debugging bugs
and regressions.
Immutable samplers requires that the samplers stay... immutable, hence
this boolean is useful if the promise gets broken. We might want to turn
this into a `GLOBAL_DEF` setting.
Instead of creating dozen/hundreds/thousands of `VkDescriptorSet` every
frame that need to be freed individually when they are no longer needed,
they all get freed at once by resetting the whole pool. Once the whole
pool is no longer in use by the GPU, it gets reset and its memory
recycled. Descriptor sets that are created to be kept around for longer
or forever (i.e. not created and freed within the same frame) **must
not** use linear pools. There may be more than one pool per frame. How
many pools per frame Godot ends up with depends on its capacity, and
that is controlled by
`rendering/rendering_device/vulkan/max_descriptors_per_pool`.
- **Possible improvement for later:** It should be possible for Godot
to adapt to how many descriptors per pool are needed on a per-key basis
(i.e. grow their capacity like `std::vector` does) after rendering a few
frames; which would be better than the current solution of having a
single global value for all pools (`max_descriptors_per_pool`) that the
user needs to tweak.
- `bool linear_descriptor_pools_enabled = true`
Setting it to false enforces the old behavior. Useful for debugging bugs
and regressions.
Setting it to false is required when workarounding driver bugs (e.g.
Adreno 730).
A ridiculous optimization. Ridiculous because the original code
should've done this in the first place. Previously Godot was doing the
following:
1. Create a command buffer **pool**. One per frame.
2. Create multiple command buffers from the pool in point 1.
3. Call `vkBeginCommandBuffer` on the cmd buffer in point 2. This
resets the cmd buffer because Godot requests the
`VK_COMMAND_POOL_CREATE_RESET_COMMAND_BUFFER_BIT` flag.
4. Add commands to the cmd buffers from point 2.
5. Submit those commands.
6. On frame N + 2, recycle the buffer pool and cmd buffers from pt 1 &
2, and repeat from step 3.
The problem here is that step 3 resets each command buffer individually.
Initially Godot used to have 1 cmd buffer per pool, thus the impact is
very low.
But not anymore (specially with Adreno workarounds to force splitting
compute dispatches into a new cmd buffer, more on this later). However
Godot keeps around a very low amount of command buffers per frame.
The recommended method is to reset the whole pool, to reset all cmd
buffers at once. Hence the new steps would be:
1. Create a command buffer **pool**. One per frame.
2. Create multiple command buffers from the pool in point 1.
3. Call `vkBeginCommandBuffer` on the cmd buffer in point 2, which is
already reset/empty (see step 6).
4. Add commands to the cmd buffers from point 2.
5. Submit those commands.
6. On frame N + 2, recycle the buffer pool and cmd buffers from pt 1 &
2, call `vkResetCommandPool` and repeat from step 3.
**Possible issues:** @dariosamo added `transfer_worker` which creates a
command buffer pool:
```cpp
transfer_worker->command_pool =
driver->command_pool_create(transfer_queue_family,
RDD::COMMAND_BUFFER_TYPE_PRIMARY);
```
As expected, validation was complaining that command buffers were being
reused without being reset (that's good, we now know Validation Layers
will warn us of wrong use).
I fixed it by adding:
```cpp
void RenderingDevice::_wait_for_transfer_worker(TransferWorker
*p_transfer_worker) {
driver->fence_wait(p_transfer_worker->command_fence);
driver->command_pool_reset(p_transfer_worker->command_pool); //
! New line !
```
**Secondary cmd buffers are subject to the same issue but I didn't alter
them. I talked this with Dario and he is aware of this.**
Secondary cmd buffers are currently disabled due to other issues (it's
disabled on master).
- `bool RenderingDeviceCommons::command_pool_reset_enabled`
Setting it to false enforces the old behavior. Useful for debugging bugs
and regressions.
There's no other reason for this boolean. Possibly once it becomes well
tested, the boolean could be removed entirely.
Adds `command_bind_render_uniform_sets` and
`add_draw_list_bind_uniform_sets` (+ compute variants).
It performs the same as `add_draw_list_bind_uniform_set` (notice
singular vs plural), but on multiple consecutive uniform sets, thus
reducing graph and draw call overhead.
- `bool descriptor_set_batching = true;`
Setting it to false enforces the old behavior. Useful for debugging bugs
and regressions.
There's no other reason for this boolean. Possibly once it becomes well
tested, the boolean could be removed entirely.
Godot currently does the following:
1. Fill the entire cmd buffer with commands.
2. `submit()`
- Wait with a semaphore for the swapchain.
- Trigger a semaphore to indicate when we're done (so the swapchain
can submit).
3. `present()`
The optimization opportunity here is that 95% of Godot's rendering is
done offscreen.
Then a fullscreen pass copies everything to the swapchain. Godot doesn't
practically render directly to the swapchain.
The problem with this is that the GPU has to wait for the swapchain to
be released **to start anything**, when we could start *much earlier*.
Only the final blit pass must wait for the swapchain.
TheForge changed it to the following (more complicated, I'm simplifying
the idea):
1. Fill the entire cmd buffer with commands.
2. In `screen_prepare_for_drawing` do `submit()`
- There are no semaphore waits for the swapchain.
- Trigger a semaphore to indicate when we're done.
3. Fill a new cmd buffer that only does the final blit to the
swapchain.
4. `submit()`
- Wait with a semaphore for the submit() from step 2.
- Wait with a semaphore for the swapchain (so the swapchain can
submit).
- Trigger a semaphore to indicate when we're done (so the swapchain
can submit).
5. `present()`
Dario discovered this problem independently while working on a different
platform.
**However TheForge's solution had to be rewritten from scratch:** The
complexity to achieve the solution was high and quite difficult to
maintain with the way Godot works now (after Übershaders PR).
But on the other hand, re-implementing the solution became much simpler
because Dario already had to do something similar: To fix an Adreno 730
driver bug, he had to implement splitting command buffers. **This is
exactly what we need!**. Thus it was re-written using this existing
functionality for a new purpose.
To achieve this, I added a new argument, `bool p_split_cmd_buffer`, to
`RenderingDeviceGraph::add_draw_list_begin`, which is only set to true
by `RenderingDevice::draw_list_begin_for_screen`.
The graph will split the draw list into its own command buffer.
- `bool split_swapchain_into_its_own_cmd_buffer = true;`
Setting it to false enforces the old behavior. This might be necessary
for consoles which follow an alternate solution to the same problem.
If not, then we should consider removing it.
PR #90993 added `shader_destroy_modules()` but it was not actually in
use.
This PR adds several places where `shader_destroy_modules()` is called
after initialization to free up memory of SPIR-V structures that are no
longer needed.
This fixes a bug in the analyzer where it did not push the SHADOWED_VARIABLE_BASE_CLASS
warning for members shadowed by variable in subclass. It does this by comparing the class
which contains the shadowed member with the class containing the variable, and pushing
SHADOWED_VARIABLE only if the classes are the same. Additionally, SHADOWED_VARIABLE_BASE_CLASS
can take an extra symbol which helps to specify the line for non native base class.
- Adds Swappy for Android for stable frame pacing
- Implements pre-transformed Swapchain so that Godot's compositor is in
charge of rotating the screen instead of Android's compositor
(performance optimization for phones that don't have HW rotator)
============================
The work was performed by collaboration of TheForge and Google. I am
merely splitting it up into smaller PRs and cleaning it up.
Changes from original PR:
- Removed "display/window/frame_pacing/android/target_frame_rate" option
to use Engine::get_max_fps instead.
- Target framerate can be changed at runtime using Engine::set_max_fps.
- Swappy is enabled by default.
- Added documentation.
- enable_auto_swap setting is replaced with swappy_mode.
The project setting does not reflect CLI argument overrides
(including `--headless` which sets the audio driver to `Dummy`),
so it can't be reliably used to detect which audio driver is
actually being used at run-time.
This adds support in all backends, but the Compatibility renderer works the best.
Mobile and Forward+ can only support one directional light shader (the first in the tree)
While the Compatibility renderer supports any number of shadows.
Co-authored-by: Clay John <claynjohn@gmail.com>
If the module is enabled (default), 2D physics works as it did before.
If the module is disabled and no other 2D physics server is registered
(via a module or GDExtension), then we fall back to a dummy
implementation which effectively disables 2D physics functionality (and
a warning is printed).
The dummy 2D physics server can also be selected explicitly, in which
case no warning is printed.
If the module is enabled (default), 3D physics works as it did before.
If the module is disabled and no other 3D physics server is registered
(via a module or GDExtension), then we fall back to a dummy
implementation which effectively disables 3D physics functionality (and
a warning is printed).
The dummy 3D physics server can also be selected explicitly, in which
case no warning is printed.
- Returns an empty list when there's not registered plugins, thus preventing the creation of spurious iterator objects
- Inline `Godot#getRotatedValues(...)` given it only had a single caller. This allows to remove the allocation of a float array on each call and replace it with float variables
- Disable sensor events by default. Sensor events can fired at 10-100s Hz taking cpu and memory resources. Now the use of sensor data is behind a project setting allowing projects that have use of it to enable it, while other projects don't pay the cost for a feature they don't use
- Create a pool of specialized input `Runnable` objects to prevent spurious, unbounded `Runnable` allocations
- Disable showing the boot logo for Android XR projects
- Delete locale references of jni strings
- Improve documentation related to embedded subwindows and single-window mode.
- Add `minsize` keyword aliases for `popup_centered_clamped()` methods
to ease migration to Godot 4.x (this was Godot 3.x terminology).
Adds 3D fixed timestep interpolation to the rendering server.
This does not yet include support for multimeshes or particles.
Co-authored-by: lawnjelly <lawnjelly@gmail.com>
- Add support for dispatching input on the render thread (UI thread is the current default) when `input_buffering` and `accumulated_input` are disabled. At the expense of latency, this helps prevent 'heavy' applications / games from blocking the UI thread (the default behavior) which may cause the application to ANR.
- Remove GLSurfaceView logic causing the UI thread to wait on the GL thread during lifecycle events. The removed logic would cause the UI thread to ANR when the GL thread is blocked.
https://github.com/godotengine/godot/pull/83976 added support for 3D MSAA on the Compatibility renderer, but it also mentions that on platforms other than Android, foveated rendering will not work if MSAA is enabled.
I removed the note saying that 3D MSAA is not supported on compatibility and added a comment mentioning that foveated rendering won't work on platforms other than android if 3D MSAA is enabled and added the alternative/equivalent for desktop.
Folder names ending with one or more `.` characters are not allowed
on Windows, so this would break writing logs, shader cache and other
project-specific files. Trailing periods are now stripped in this case.
On non-Windows platforms, this change still applies in the interest
of portability.
This preserves compatibility when upgrading Godot 4.2 projects which relied on that
path being configured in the editor settings.
The old name also makes sense for this one, it's fine for fbx2gltf_path to be under
a generic fbx category which could have more settings also impacting ufbx.
Disabling V-Sync when using the Compatibility rendering method works
since Godot 4.2 at least.
This also documents the `--disable-vsync` command line argument.
Due to the low resolution of the occlusion buffer, small gaps between occluders can be closed and incorrectly occlude instances which should show through the gaps. To ameliorate this problem, this PR jitters the occlusion buffer over time, making it more likely an instance will be seen through a gap. This is used in conjunction with an occlusion timer per instance, to prevent instances flickering on and off rapidly.
Adds `ui_text_skip_selection_for_next_occurrence` action and related implementation to text editor.
This action is bound `Ctrl+Alt+D` shorcut.
Used in conjonction with `ui_add_skip_selection_for_next_occurrence`, it gives the user the ability to select many occurrences of a selection
and avoid some of them.
Used without a previous selection, the action jumps to the next occurrence of the current word under the caret.
Adds fixed timestep interpolation to the rendering server (2D only).
Switchable on and off with a project setting (default is off).
Co-authored-by: lawnjelly <lawnjelly@gmail.com>
Defaults to "Auto", which detects the casing based on the
preference of the currently selected language (C# for example
prefers PascalCase whereas GDScript prefers snake_case).
This update introduces a new import method for FBX files using ufbx. If the fbx2gltf import fails, it will use the most recently cached scene from the ufbx import. The process is sped up by introducing threads to load the ufbx portion.
Key changes include:
- Support for importing geometry helper nodes in FBX files.
- Addition of cameras and lights with updated names.
- Removal of the fbx importer manager.
- Introduction of ModelDocument3D and updates to its methods.
- Changes to FBX import options and visibility.
- Updating the documentation and handling some errors.
- Store the original non-unique node, mesh and animation names in FBX and glTF.
Co-Authored-By: bqqbarbhg <bqqbarbhg@gmail.com>
This small in-engine documentation change aims to make it easier to
discover ways to handle pixel art aesthetics.
- I have moved 2D pixel snapping settings out of "Advanced." This now matches other pixel-art-friendly settings for
GUI Snapping and Default Texture filtering.
- I've added notes to the project settings and Sprite/AnimatedSprite sources to hint users towards better understanding of why pixelated sprites may not work correctly and what to do about it. This should help users make informed decisions for their needs.
Context: Proper handling of pixel art in Godot is routinely frustrating for new users: I, like others, assumed that Godot would act on pixels, not subpixels, when I was working a pixel art game. I was confused when my interpolations would appear blurry, and when pixel textures would be distorted for no apparent reason (this was because of centering).
I had naively thought that setting Linear interpolation would be the single "it's a pixel art game" toggle, but that only hid the underlying issues until later. I had no idea there was a snap-to-pixel option because it was hidden in the Advanced options, since my default assumption was that a pixel art game would want no subpixels at all.
Some references for the frustration:
- https://github.com/godotengine/godot/issues/82696
- https://www.reddit.com/r/godot/comments/fah25e/best_way_to_achieve_pixel_perfect_rendering/
- https://shaggydev.com/2021/09/21/project-setup-for-pixel-art/
* Do not print empty line when header is disabled
* Do not print Vulcan header
* Also add "Print header" project setting (default On)
(suggested by @kaissouDev)
* Add docs for the project setting
(with suggestions by @Mickeon and @akien-mga)
Co-authored-by: Micky <66727710+Mickeon@users.noreply.github.com>
Co-authored-by: A Thousand Ships <96648715+AThousandShips@users.noreply.github.com>
Co-authored-by: Rémi Verschelde <rverschelde@gmail.com>
The editor setting makes it possible to tweak V-Sync status
independently of the project setting.
Use cases:
- Decrease input lag and increase editor responsiveness when editing
a project that has V-Sync enabled.
- Avoid tearing when editing a project that has V-Sync disabled.
Not everything is yet implemented, either for Godot or personal
limitations (I don't have all hardware in the world). A brief list of
the most important issues follows:
- Single-window only: the `DisplayServer` API doesn't expose enough
information for properly creating XDG shell windows.
- Very dumb rendering loop: this is very complicated, just know that
the low consumption mode is forced to 2000 Hz and some clever hacks are
in place to overcome a specific Wayland limitation. This will be
improved to the extent possible both downstream and upstream.
- Features to implement yet: IME, touch input, native file dialog,
drawing tablet (commented out due to a refactor), screen recording.
- Mouse passthrough can't be implement through a poly API, we need a
rect-based one.
- The cursor doesn't yet support fractional scaling.
- Auto scale is rounded up when using fractional scaling as we don't
have a per-window scale query API (basically we need
`DisplayServer::window_get_scale`).
- Building with `x11=no wayland=yes opengl=yes openxr=yes` fails.
This also adds a new project property and editor setting for selecting the
default DisplayServer to start, to allow this backend to start first in
exported projects (X11 is still the default for now). The editor setting
always overrides the project setting.
Special thanks to Drew Devault, toger5, Sebastian Krzyszkowiak, Leandro
Benedet Garcia, Subhransu, Yury Zhuravlev and Mara Huldra.
Existing shadow caster culling takes no account of the camera.
This PR adds the highly encapsulated class RenderingLightCuller which can cut down the casters in the shadow volume to only those which can cast shadows on the camera frustum.
This works even if file logging is disabled in the project settings,
or for the editor/project manager.
`--log-file`'s value can be an absolute path or relative to the project
directory (similar to existing arguments like `--write-movie`).
Adds a new system to automatically reorder commands, perform layout transitions and insert synchronization barriers based on the commands issued to RenderingDevice.
This commits rewrites the sync logic in a way that the
`use_system_threads_for_low_priority_tasks` setting, which was added due to
the lack of a cross-platform wait-for-multiple-objects functionality, can be
removed (it's as if it was effectively hardcoded to `false`).
With the new implementation, we have the best of both worlds: threads don't
have to poll, plus no bespoke threads are used.
In addition, regarding deadlock prevention, since not every possible case of
wait-deadlock could be avoided, this commits removes the current best-effort
avoidance mechanisms and keeps only a simple, pessimistic way of detection.
It turns out that the only current user of deadlock prevention, ResourceLoader,
works fine with it and so every possible situation in resource loading is now
properly handled, with no possibilities of deadlocking. There's a comment in
the code with further details.
Lastly, a potential for load tasks never being awaited/disposed is cleared.
Fixes#82587 by improving usability of the untyped_declaration warning. This adds a note to the documentation that recommends turning on EditorSettings.text_editor/completion/add_type_hints when the untyped declaration warning is set to warn or error.
Co-authored-by: Yuri Sizov <11782833+YuriSizov@users.noreply.github.com>
Port over the logic from Godot 3.5 for indirect lighting. This should fix many issues about indirect bounces causing more energy and improve the overall quality of the result.
- Provide api to retrieve the running Godot instance from a GodotHost
- Provide api for the GodotHost to register runtime GodotPlugin instances
- Hide the GodotService class until it's completed
- Include project setting to enable long press for Android devices
- Include project setting to enable pan and scale gestures on Android devices
This commit also adds means to manually disable warnings
in `code` tags where it's a false positive with the new
`skip-lint` attribute.
Warnings are now enabled on CI to prevent future errors.
[macOS] Use ANGLE by default if engine is built with statically linked ANGLE.
[Windows] Add option enable/disable fallback to ANGLE and option force specific GPU to always use ANGLE.
This means no CPU occlusion culling (and not compiling Embree), unless
you compile custom export templates with `module_raycast_enabled=yes`.
This reduces the memory footprint significantly, and binary size.
Fixes#70621.
Co-authored-by: Hugo Locurcio <hugo.locurcio@hugo.pro>
Hendrik Brucker
Rémi Verschelde
Thaddeus Crews
Bastiaan Olij
smix8
Dario
Rémi Verschelde
Mikael Hermansson
Matias N. Goldberg
clayjohn
BlueCube3310
tetrapod00
Wagner
Micky
girdenis-p
Stuart Carnie
Hugo Locurcio
Max Hilbrunner
ywmaa
Ricardo Buring
A Thousand Ships
SheepYhangCN
bruvzg
David Snopek
Programneer
Wierdox
Fredia Huya-Kouadio
kobewi
Kusok
jsjtxietian
Eric allen
Danil Alexeev
Daniel Castellanos
Adam Scott
Marcel Offermans
skyace65
George Marques
Haoyu Qiu
Manpreet Singh
lawnjelly
Christophe Andral
Pawel Lampe
RedMser
K. S. Ernest (iFire) Lee
Sabs, like Sobs
Anatoli Babenia
Silc Lizard (Tokage) Renew
ALEKSEY KAPUSTYANENKO
Tiger Jove
Riteo
mxaddict
Yuri Sizov
MewPurPur
Erik Johnson
Pedro J. Estébanez
Septian
Allen Pestaluky
Georg Wacker
Fredia Huya-Kouadio
Jordyfel
ryanabx