Lines loose extraction cannot happen on the fly as the render data isn't
complete. This patch will do the lines loose subbuffer extraction as a
simple task and not on the fly.
This patch introduces 2 sub-types of tasks that are executed single threaded.
`EXTRACT_MESH_EXTRACT` would do the extraction using Mesh/BMesh geometry.
`EXTRACT_LINES_LOOSE` creates the `lines_loose` subbuffer from already cached
`lines` IBO.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7964
The underlying issue is that the loose edge extraction cannot be
threaded. This patch will extract the loose edges during initialization
of the render mesh data.
When working on this patch the mesh_render_data_update was split into
multiple functions to identify what part was failing. These functions
would also help us with debugging.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7962
This patch uses a graph flow scheduler for creating all mesh batches.
On a Ryzen 1700 the framerate of Spring/020_02A.anim.blend went from 10 fps to 11.5 fps.
For each mesh where batches needs to be updated a sub-graph will be added to the task_graph.
This sub-graph starts with an extract_render_data_node. This fills/converts the required data
from Mesh.
Small extractions and extractions that can't be multi-threaded are grouped in a single
`extract_single_threaded_task_node`.
Other extractions will create a node for each loop exceeding 4096 items. these nodes are
linked to the `user_data_init_task_node`. the `user_data_init_task_node` prepares the userdata
needed for the extraction based on the data extracted from the mesh.
Note: If the `lines` and `lines_loose` are requested, the `lines_loose` sub-buffer is created
as part of the lines extraction. When the lines_loose is only requested the sub-buffer is
created from the existing `lines` buffer. It is assumed that the lines buffer is always
requested before or together with the lines_loose what is always the case (see
`DRW_batch_requested(cache->batch.loose_edges, GPU_PRIM_LINES)` in `draw_cache_impl_mesh.c`).
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7618
This resolves a performance regression in 2.8x where every edit-mode
update performed an edit-mesh to mesh conversion.
Now the conversion will be lazily initialized if/when it's required.
New BKE_mesh_wrapper_* functions abstract over mesh data access.
Currently only edit-mesh and regular meshes are supported.
In the future sub-surface meshes may be supported too.
This patch enables TBB as the default task scheduler. TBB stands for Threading Building Blocks and is developed by Intel. The library contains several threading patters. This patch maps blenders BLI_task_* function to their counterpart. After this patch we can add more patterns. A promising one is TBB:graph that can be used for depsgraph, draw manager and compositor.
Performance changes depends on the actual hardware. It was tested on different hardwares from laptops to workstations and we didn't detected any downgrade of the performance.
* Linux Xeon E5-2699 v4 got FPS boost from 12 to 17 using Spring's 04_010_A.anim.blend.
* AMD Ryzen Threadripper 2990WX 32-Core Animation playback goes from 9.5-10.5 FPS to 13.0-14.0 FPS on Agent 327 , 10_03_B.anim.blend.
Reviewed By: brecht, sergey
Differential Revision: https://developer.blender.org/D7475
Caused by rBe82827bf6ed5.
DRW_draw_depth_object calls DRW_mesh_batch_cache_create_requested with
NULL scene, but that is accessed later on...
Scene is actually available, so pass that around.
Maniphest Tasks: T76131
Differential Revision: https://developer.blender.org/D7540
The loop normal VBO is used in two manners. In edit mode to draw the
edge normals. And in paint mode to draw the wireframe. This commit
checks which VBO is needed and build the correct one.
This allows show the wireframe correct in paint mode, when the object is
subdivided.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7419
Tasks: move priority from task to task pool {rBf7c18df4f599fe39ffc914e645e504fcdbee8636}
Tasks: split task.c into task_pool.cc and task_iterator.c {rB4ada1d267749931ca934a74b14a82479bcaa92e0}
Differential Revision: https://developer.blender.org/D7385
Since rBcf258b02f449, only wires and edges that are mapped to the
original mesh were drawn if the mesh was modified by modifiers.
Above commit was only meant for showing orig wires for paint mask
overlays [where final wireframe is not desired], so now only use
MR_EXTRACT_MAPPED when we are in a paint mode.
Maniphest Tasks: T75343
Differential Revision: https://developer.blender.org/D7333
Draw batch extraction wrongly assumed that when mapped extraction
happened that all original data could be found. This is not the case as
mapped extraction is also enabled when part of the data is present.
This fix does additional nullptr checks.
When using generative modifiers too many center dots were rendered in
the normal overlay. This patch only renders the normals of original
center dots.
Known issue: decoding the `norAndFlag` has issues on Intel GPU.
The loop normals were always drawn. We used to only draw the normals if
it was mapped to an original loop of the mesh. Due to recent changes we
can not find the correct loop and decide if we need to draw them.
Note still need to check the face dots normals. This is more complicated
as facedot normals needs to be encoded in a different way
The paint mask overlay showed the wires and edges of the final mesh.
This change will only draw wires and edges that are mapped to the
original mesh.
This change enables mapping data in regular Mesh extraction. This
can also be used for better drawing of the normal overlay.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D7277
When in editmode faces can be hidden, but in object mode these faces are
still visible. The flag if a face was hidden in edit mode is stored in
object mode, but should not be used.
The edge detection gpu batch did detect hidden faces and didn't add them
to the draw batch. The edge detection gpu batch is used for workbench
shadows, custom bone shapes and object outlines.
This patch adds all faces to the edge detection batch.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D7157
This check box alters how weights are displayed and painted,
similar to Multi Paint, but in a different way. Specifically,
weights are presented as if all locked vertex groups were
deleted, and the remaining deform groups normalized.
The new feature is intended for use when balancing weights within
a group of bones while all others are locked. Enabling the option
presents weight as if the locked bones didn't exist, and their
weight was proportionally redistributed to the editable bones.
Conversely, the Multi-Paint feature allows balancing a group of
bones as a whole against all unselected bones, while ignoring
weight distribution within the selected group.
This mode also allows temporarily viewing non-normalized weights
as if they were normalized, without actually changing the values.
Differential Revision: https://developer.blender.org/D3837
- Use 'BKE_object_defgroup' prefix for object functions.
- Rename 'defvert_verify_index' to 'defvert_ensure_index'
since this adds the group if it isn't found.
This patch adds a dedicated path to extract 16bit normals instead of packing them into 10bits/comp.
The tangents are also packed to 10bits/comp if not using the new High Quality Normal option.
Fix T61024 Degraded texture shading on dense meshes
Reviewed By: brecht
Differential Revision: https://developer.blender.org/D6614
This change will not render the normals for faces that are hidden.
Before we had instance drawing the hidden faces were registered in the
index buffer. During the overlay refactoring the rendering was migrated
to instance rendering. Instance rendering does not use the index buffer
so the data was ignored.
This patch stored the normal visibility in the .w part of the normal or
for face normals it will set the normal to zero. The shader looks at
this and renders the normals fully transparent when detected.
Reviewed By: Clément Foucault
Differential Revision: https://developer.blender.org/D6798
This way we remove the need for the srgb boolean uniform and a lot of code complexity. However, mesh update is going to be a bit slower.
I did not benchmark the performance impact.
This also fix a typo in draw_cache_impl_particles.c and fix hair not using vertex color in workbench.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D6610
The `in int flag;` in `gpu_shader_2D_edituvs_faces_vert.glsl`
don't have the values `FACE_UV_ACTIVE` and `FACE_UV_SELECT`.
Add face flags then.
Original patch is from @EitanSomething
Differential revision: https://developer.blender.org/D6520
In a recent refactor we splitted the lines extractor in `extract_lines` and
`extract_lines_loose`. When an object is in edit mode the extracted
lines loose also had to include a dummy bmesh edge iterator. This change
adds this missing dummy method.
Reviewed By: antoniov
Differential Revision: https://developer.blender.org/D6499
The ratio for area stretching was packed into an unsigned int, but could
contain negative numbers. This flipped the negative numbers to high
positive numbers and rendered the wrong color in the stretching overlay.
I can remember during {T63755} I had to flip the sign to get the
correct result, but couldn't find out why that was needed. Now I know.
Reviewed By: fclem, mano-wii
Differential Revision: https://developer.blender.org/D6440
Introduced by 9c337fcfe2 mistaken that `MeshExtract.use_threading` set
to false means that no threading is used at all. This is not the case it
will still perform threading for large objects, it will only run the
different subtasks in serial.
Fixed by scheduling the `lines_loose` in the task_pool after the rest
have been executed. This is cleaner than the previous implementation as
it sticks more to the actual design.
The lines index buffer can contain all edges (edit mode) or only loose
edges (object mode). When switching between these modes the wrong
content of the index buffer can be used.
This patch will clear the lines index buffer when a `loose_edges` is requested. Making sure it is always up to date.
Note that this is supporting an exising hack where the IBO is truncated
during the creation. We should find a different way how to solve these
kind of issues.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D6349
This is the unification of all overlays into one overlay engine as described in T65347.
I went over all the code making it more future proof with less hacks and removing old / not relevent parts.
Goals / Acheivements:
- Remove internal shader usage (only drw shaders)
- Remove viewportSize and viewportSizeInv and put them in gloabl ubo
- Fixed some drawing issues: Missing probe option and Missing Alt+B clipping of some shader
- Remove old (legacy) shaders dependancy (not using view UBO).
- Less shader variation (less compilation time at first load and less patching needed for vulkan)
- removed some geom shaders when I could
- Remove static e_data (except shaders storage where it is OK)
- Clear the way to fix some anoying limitations (dithered transparency, background image compositing etc...)
- Wireframe drawing now uses the same batching capabilities as workbench & eevee (indirect drawing).
- Reduced complexity, removed ~3000 Lines of code in draw (also removed a lot of unused shader in GPU).
- Post AA to avoid complexity and cost of MSAA.
Remaining issues:
- ~~Armature edits, overlay toggles, (... others?) are not refreshing viewport after AA is complete~~
- FXAA is not the best for wires, maybe investigate SMAA
- Maybe do something more temporally stable for AA.
- ~~Paint overlays are not working with AA.~~
- ~~infront objects are difficult to select.~~
- ~~the infront wires sometimes goes through they solid counterpart (missing clear maybe?) (toggle overlays on-off when using infront+wireframe overlay in solid shading)~~
Note: I made some decision to change slightly the appearance of some objects to simplify their drawing. Namely the empty arrows end (which is now hollow/wire) and distance points of the cameras/spots being done by lines.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D6296
I believed the crash I experienced happened because:
1. The `extract_pos_nor_init` function is called.
2. Tasks are added to the task pool for `extract_pos_nor`.
3. The tasks begin to be executed while more tasks are added.
4. In some rare cases, all existing tasks are finished, but not all have been added yet.
5. This let the task-counter go down to zero.
6. This triggered a call to `extract_pos_nor_finish`.
7. Then more tasks are added and in the end `extract_pos_nor_finish` is called again.
A solution is to use a task pool that is suspended when created.
Unfortunately, there was an outdated comment, that was probably the root cause of the issue.
Reviewers: fclem, sergey
Differential Revision: https://developer.blender.org/D5680
Support for UV Stretching overlay during multi object editing. The
VBO now holds the ratios per fase. In the shader these ratios will
be compared against the global ratios. The global rations are created
from all selected objects.
The current implementation does not fit well with the draw module. The
plan is to move the drawing of other spaces towards the draw manager what
leads to a better fit. Currently the details on this solution is unclear
but this requirement will become an attentionpoint in the future design.
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D5665
Calculating UV Stretching on large meshes showed garbage. The reason
is that the calculation is not thread save. Temporarily disable
threading for UV Stretching
