This change adds timeline semaphores to track submissions. The previous
implementation used a fence.
Timeline semaphores can be tracked in more detail as it is an counter.
For each submission the counter can be stored locally and when waiting
for completion the counter can be retrieved again and checked if is
known to be succeeded by a higher value.
The timeline semaphore is stored next to the queue and can also be used
to synchronize between multiple contexts.
Pull Request: https://projects.blender.org/blender/blender/pulls/115357
Texture usage flag `GPU_TEXTURE_USAGE_MIP_SWIZZLE_VIEW`
was originally implemented and used too conservatively for many
cases in which the underlying API flags were not required.
Renaming to `GPU_TEXTURE_USAGE_FORMAT_VIEW` to reflect
the only essential use case for when a texture view is initialized with
a different texture format to the source texture. Texture views can
still be created without this flag when mip range or base level is
adjusted,
This flag is still required by stencil views and internally by the Metal
backend for certain feature support such as SRGB render toggling.
Patch also includes some small changes to the Metal backend to
adapt to this new compatibility and correctly capture all texture view
use-cases.
Related to #115269
Authored by Apple: Michael Parkin-White
Pull Request: https://projects.blender.org/blender/blender/pulls/115300
Move tagging from `flush_editors_id_update` to `deg::id_tag_update`.
Since ID relationships cannot always be trusted when tagging for update
happens (can happen e.g. during liboverride apply itself, when deleting
IDs, and so on), embedded IDs are simply tagged here too, instead of
directly tagging their owner ID.
Propagation from embedded to owner ID then happens when checking and
'consuming' the `LIB_TAG_LIBOVERRIDE_AUTOREFRESH` ID tag, in
`BKE_lib_override_library_main_operations_create`.
Pull Request: https://projects.blender.org/blender/blender/pulls/115319
Currently all buffer types were stored in host memory, which is visible to the GPU as well.
This is typically slow as the data would be transferred over the PCI bus when used.
Most of the time Index and Vertex buffers are written once and read many times so it makes
more sense to locate them on the GPU. Storage buffers typically require quick access as they
are created for shading/compute purposes.
This PR will try to store vertex buffers, index buffers and storage buffers on device memory
to improve the performance.
Uniform buffers are still located on host memory as they can be uploaded during binding process.
This can (will) reset the graphics pipeline triggering draw calls using unattached resources.
In future this could be optimized further as in:
* using different pools for allocating specific buffers, with a fallback when buffers cannot be
stored on the GPU anymore.
* store uniform buffers in device memory
Pull Request: https://projects.blender.org/blender/blender/pulls/115343
This restrict the usage of distant light if the world shader
contains absorption phenomenon.
Why this is needed is described in #114062.
At the same time, this removes the parameter for enabling
volume light as this is now an auto-detected optimization.
This also contains a few small cleanups.
Pull Request: https://projects.blender.org/blender/blender/pulls/115284
This PR adds the default vulkan pipeline cache to when creating
pipelines.
Pipeline caching reuses pipelines that have already been created.
Pipeline creation can be somewhat costly - it has to compile the
shaders at creation time for example.
The cache is recreated at each run of Blender.
Pull Request: https://projects.blender.org/blender/blender/pulls/115346
Some test cases inside the vulkan backend don't rely on an initialized
Vulkan Context and can always be run.
This PR enables those test cases when `WITH_GTEST` and
`WITH_VULKAN_BACKEND` is on. Allowing some tests to run on the
buildbot.
Pull Request: https://projects.blender.org/blender/blender/pulls/114574
When pressing `I` in the viewport, the colors of the FCurve channels were no longer set correctly.
Caused by #113504
Fix by removing the flag that determined that in the first place,
and read straight from the user preferences. Then move the code
that sets the mode on the FCurve into the function that actually creates it.
For 99% of cases the code went to the user preference flag `AUTOKEY_FLAG_XYZ2RGB`
and if that was set, the `INSERTKEY_XYZ2RGB` would be set. The only case where this
was not from the user preferences was from custom keying sets.
There was an override flag for FCurve colors on custom keying sets.
I removed that with this patch since the use case is not apparent
and custom keying sets are hardly used.
Pull Request: https://projects.blender.org/blender/blender/pulls/115297
The problem here was that `RNA_parameter_set_lookup` was used incorrectly.
Instead of passing in the `PointerRNA`, one can pass in the value directly. If one
wanted to pass the `PointerRNA`, one would have to use `PARM_RNAPTR`, but
that's not necessary here.
Pull Request: https://projects.blender.org/blender/blender/pulls/115347
vcpkg somewhat forcefully injects itself into the build system
causing some work stations to use the vcpkg headers/libs rather than
the headers from our lib folder,we added some mitigation against this
in 34b3a9583a but still the occasional
user with build issues showed up, this should nip that in the bud once
and for all.
Harfbuzz and FriBiDi are included in our external libraries for all
platforms. This PR adds the glue to make them available as optional
build components (off by default).
Pull Request: https://projects.blender.org/blender/blender/pulls/114947
# Issue
Having a lot of keys in your scene can dramatically slow down the Dope Sheet.
That is because everytime the Dope Sheet is redrawn,
the Keylists have to be recomputed.
In the case of the summary channel, that means going through
all keyframes of all the `FCurves` and adding them to the keylist.
That eats up 95% of the time it takes to draw a frame.
# This PR
It's not a perfect solution, rather it solves the performance issue
for the case when you are not displaying all keys.
Instead of going through all the keys, only add the
keys visible in the view to the keylist.
This speeds up the Dope Sheet significantly
depending on the zoom level.
This also improves the responsiveness when selecting and transforming keyframes.
# Performance changes
The function measured is `ED_channel_list_flush`
which is responsible for building the keylists and drawing them.
The test setup contains 62 bones with all
10 channels keyed (location, rot quaternion, scale) on 6000 frames.
So 3.720.000 keys. The heavier the dataset the bigger the performance impact.
The data was recorded with only the Dope Sheet open, and 3 channels visible
* Summary
* Object
* Action
The more channels are visible, the greater the performance gain. This can be seen in the video.
| visible range | before | after |
| - | - | - |
| 200f | 250ms | 10ms |
| 400f | 250ms | 18ms |
| 3000f | 250ms | 130ms |
| 6000f | 250ms | 250ms |
Pull Request: https://projects.blender.org/blender/blender/pulls/114854
While it may have been working from a practical PoV (not certain though,
since some bug would prevent clearing runtime data when writing embedded
Scene collection in current code), this is semantically wrong.
The owner of an embedded ID is a critical piece of information in
Blender data structure and ID management code. Having it written in
.blend files is also a potential good source of data for investigating
issues.
Further more, this handling of `owner` ID data is somewhat generic now
in ID management, so if this data should be considered runtime, then the
change should also be made in NodeTree and Key IDs.
This commit partially reverts 44dd3308a5, in the future I'd like to
be involved in the review of changes affecting ID management.
NOTE: fix for embedded collection runtime data not being cleared on
write will be committed separately.
When in scene mode, operation searches the object in master collection
to unlink. But object can be in any collection. So iterate through all
collections present in scene to unlink the object
Pull Request: https://projects.blender.org/blender/blender/pulls/112955
Motivation: When discussing with @Jeroen-Bakker and @aras_p about how to
approach sorting when rendering Gaussian splats in Blender, we realised that
compute shaders could help there (and have many other use cases), and that also
due to Blender 4.0 being on OpenGL >= 4.3, we can now rely on compute shaders
existing.
This PR is an initial pass for that functionality. It comes with a Python example, which
runs a compute shader and saves the output to a texture, which is then rendered in the
viewport.
There is no exposed support for storage buffers yet, but I expect I'll be able to work on
them soon if this is accepted.
The newly added parts are:
1. `gpu.compute.dispatch()`
2. a way set the compute source to `GPUShaderCreateInfo`: `shader_info.compute_source()`
3. a way to set the image store for `GPUShaderCreateInfo`: `shader_info.image()`
4. a way to set the `local_group_size` for `GPUShaderCreateInfo`: `shader_info.local_group_size(x,y,z)`
5. a way to get `max_work_group_size` from capabilities: `gpu.capabilities.max_work_group_size_get(index)`
6. a way to get `max_work_group_count` from capabilities: `gpu.capabilities.max_work_group_count_get(index)`
Pull Request: https://projects.blender.org/blender/blender/pulls/114238
OpenImageDenoise has two modes, high quality and balanced. This now exposes the modes as user parameters, with viewport denoising defaulting to balanced and final frame rendering set to high quality.
Ref #115045
Co-authored-by: Werner, Stefan <stefan.werner@intel.com>
Pull Request: https://projects.blender.org/blender/blender/pulls/115265
No functional changes.
Rename the `eAutokey_Flag` to `eKeyInsert_Flag`
to indicate that it is not only used for auto keying.
Also rename the enum items to better reflect what they are used for.
Pull Request: https://projects.blender.org/blender/blender/pulls/115295
These drivers crash on startup caused by a driver bug. This include the
latest drivers for legacy Intel CPUs with a HD 4000/HD 5000 series GPU.
To be on the safe side all drivers with version 20.19.15.51* will be marked
unsupported as we don't have the platforms to identify the precise driver
versions that fail.
See #113124 for more information.
Pull Request: https://projects.blender.org/blender/blender/pulls/115228
Since b1526dd2c6, viewport renderer sets `G.is_rendering`, but VSE scene
rendering function used this to decide whether to do offscreen opengl
render or use render API. This logic was quite weak.
Use `SeqRenderData::for_render` instead of `G.is_rendering`, since it
explicitly defines whether strip rendering is invoked by F12 render job.
Since offscreen gl rendering rely on `BLI_thread_is_main()` being
true, use this to initialize `do_seq_gl` variable for clarity.
The use of render job path was further conditioned on `G.is_rendering`,
with exception of running headless, but this condition was incorrect.
This condition was reformulated as precondition, which if true, returns
`nullptr` instead of crashing.
Pull Request: https://projects.blender.org/blender/blender/pulls/115079