Cleanup to avoid unnecessary copies of VArray. This
requires ref-qualifier overloads of dereference operator
of attribute reader and some move operators and constructor
overloads in the code.
Pull Request: https://projects.blender.org/blender/blender/pulls/118437
Caused by bace4c9a29 which only
extracted normals at face corners. Solve by also extracting loose edge and
loose vertex normals. And change the BMesh extraction to not use the
face iteration of the MeshExtract system, just because that's simpler.
Implements another phase of #116901, this time for the `lines` and
`lines_loose` index buffers that store indices for wireframe drawing.
The key improvement is removing loose edge's dependency on the main
edge index buffer. That means for the majority of meshes with no loose
edges, edge index extraction can be completely skipped. Even when there
are loose edges, only the loose edges need to be extracted with
wireframe turned off.
Besides that improvement, there are more changes to use data-oriented
code with visible hot loops instead of the virtual function call design
used for the existing mesh extractor system. For this step I completely
replaced the `extract_lines` object, which is in line with the general
plan for this area.
Additionally, hidden edge filtering is done ahead of time using several
`IndexMask` operations. This means only indices for visible edges need
to be uploaded to the GPU, and no restart index stripping needs to be
performed on macOS.
On my usual test file with 1.9 million vertices, I observed an
improvement from 26 to 33 FPS with wireframe off, and from 9.15 to 9.5
FPS with wireframe on.
Pull Request: https://projects.blender.org/blender/blender/pulls/120720
This requires adding a destructor and deleting move and copy assignment
for SculptSession, because (for now at least) we want to keep the PBVH
as an opaque type (though with one exception for including pbvh_intern.hh
in paint.cc for the SculptSession destructor).
Pull Request: https://projects.blender.org/blender/blender/pulls/121227
The root issue is that `shadow_punctual_footprint_ratio` was not
fed with Z distance but by radial distances to camera and
light.
This commit cleans up this computation by only precomputing the
film pixel radius on CPU. This allow the LOD computation to be
more local and easier to read.
Fix#119725
Pull Request: https://projects.blender.org/blender/blender/pulls/121167
This moves the seemingly arbitrary value of "1 / 2000"
into a constant variable
(`bke::greasepencil::LEGACY_RADIUS_CONVERSION_FACTOR`)
so that it can be used in all the places where
legacy "thickness" values need to be converted.
This also expands the explanation of the factor a bit, so it's
clearer why it is needed.
This PR implements the viewport overlay for Weight Paint mode in GPv3.
In Weight Paint mode the stroke points are colored depending on their
weights in the active vertex group.
Pull Request: https://projects.blender.org/blender/blender/pulls/118273
Compute shaders are required since 4.0. There was one occasion where
an older AMD driver failed and support was turned off. This driver
is now marked unsupported.
This PR includes:
- removing the check in viewport compositing
- remove properties from system info
- always construct draw manager.
- remove unused pass logic in draw hair/curves
- add deprecation warning when accessed from python
Pull Request: https://projects.blender.org/blender/blender/pulls/120909
Currently, `cache.eval_cache.final.attributes_buf[i] != nullptr` is used as
high level check to see if attributes need to be recomputed. However, this
buffer was not freed when a related buffer was freed that has to be
recomputed.
Move most code to `blender::bke::subdiv`. That helps organization
and makes using C++ in subdiv code easier, which will be useful for
removing the unnecessary opensubdiv C-API wrapper.
The main motivation for this is that it's part of a fix for #113377,
where I want to propagate the edit mesh pointers through copied
meshes in modifiers and geometry nodes, instead of just setting the
edit mesh pointer at the end of the modifier stack. That would have
two main benefits:
1. We avoid the need to write to the evaluated mesh, after evaluation
which means it can be shared directly among evaluated objects.
2. When an object's mesh is completely replaced by the mesh from another
object during evaluation (with the object info node), the final edit
mesh pointer will not be "wrong", allowing us to skip index-mapped
GPU data extraction.
Beyond that, using a shared pointer just makes things more automatic.
Handling of edit mesh data is already complicated enough, this way some
of the worry and complexity can be handled by RAII.
One thing to keep in mind is that the edit mesh's BMesh is still freed
manually with `EDBM_mesh_free_data` when leaving edit mode. I figured
that was a more conservative approach for now. Maybe eventually that
could be handled automatically with RAII too.
Pull Request: https://projects.blender.org/blender/blender/pulls/120276
This PR adds a context function to consider all
buffer bindings obsolete. This is in order to
track missing binds and invalid lingering states
accross `draw::Pass`es.
The functions `GPU_storagebuf_debug_unbind_all`
and `GPU_uniformbuf_debug_unbind_all` do nothing
more than resetting the internal debug slot bits
to zero. This is what OpenGL backend does as it
doesn't track the bindings themselves.
Other backends might have other way to detect
missing bindings. If not they should be
implemented separately anyway.
I renamed the function to `debug_unbind_all` to
denote that it actually does something related to
debugging.
This also add SSBO binding check for OpenGL as it
was also missing.
#### Future
This error checking logic is pretty much backend
agnostic. While it would be nice to move it at
`gpu::Context` level, we don't have the resources
for that now.
Pull Request: https://projects.blender.org/blender/blender/pulls/120716
The selection engine renders `uint` texture where each element of the
geometry is represented as an index.
When the geometry have modifiers, the original index was used for the
texture.
However this is not necessary for the select engine to work. The real
mesh indexes can be used in the texture and, only at the end, the
original index can be retrieved on the CPU itself.
Advantages of this approach:
- Optimizes the code to generate selection buffers.
- Makes the select id texture more informative as it identifies the
real elements instead of the original ones.
Pull Request: https://projects.blender.org/blender/blender/pulls/119977
Error in 0f46e02310
The maximum index in the triangles index buffer was set incorrectly,
causing the GPU code to switch to using 16 bit integers when it shouldn't.
Caused by ca0b02aa96.
The `points.size() > 1` was added because we can't access
`lengths[points.size() - 1]` if the curve has a single point.
The fix just makes sure we default to a `u` value of `0.0f`
for single point cyclic curves.
Pull Request: https://projects.blender.org/blender/blender/pulls/120565
This PR is another step in the refactor described by #116901.
This change applies to the triangle index buffer. The main improvement
is the ability to recognize when the mesh corner triangles index array
can be uploaded directly (when there is a single material and no hidden
faces). In that case the index data should be copied directly to the
GPU rather than to a temporary array owned by the IBO first. Though
that isn't implemented yet since it will be handled by the GPU module
later, the code is now structured to make that change simple from the
data extraction perspective.
Other than that, the main change is to not use the extractor iterator
framework anymore, and to set index data directly instead of using GPU
API functions. Though we're mainly bottlenecked by memory-bandwidth
anyway, it's nice to avoid function call overhead.
We also now avoid creating the array of sorted triangle indices when
there is a single material and no hidden faces. And we don't use
restart indices for the single-material case anymore. For Metal that's
nice because we can avoid `strip_restart_indices`.
I didn't notice significant performance improvements in my test files
beyond a few percent here and there. With a hacked implementation of
the copy-directly-to-the-gpu optimization, I did see more consistent
improvements though.
Pull Request: https://projects.blender.org/blender/blender/pulls/119130
Before bace4c9a29, the selection was uploaded twice: both face
and vertex selection. Afterwards only the face selection was uploaded,
but the vertex selection is necessary for the paint mode vertex overlay.
Now choose between uploading the vertex or face attribute based on the
paint overlay flags.
This adds two functions to the `Drawing` class: `fill_colors()`
and `fill_colors_for_write()`.
In places where the attribute API was used directly, these
functions are now used instead.
This is part of #119080. The point is that developers
shouldn't need to remember the name, type, default value,
and domain of built-in attributes (e.g. used by the renderer).
Pull Request: https://projects.blender.org/blender/blender/pulls/120296
Adds ".selection_handle_left" and ".selection_handle_right" attributes to
`CurvesGeometry` and their support in curve's select all, pick, box, circle
and lasso select tools.
Pull Request: https://projects.blender.org/blender/blender/pulls/119712
Implements the GPv2 onion skinning functionality.
There are no functional changes exept for the `use_ghosts_always`
option, which has been removed. This was used to show onion
skinning in the final render.
Pull Request: https://projects.blender.org/blender/blender/pulls/119792
Seems to be caused by c53e220aef. I didn't think our PBVH building would
ever create a node with no triangles, but I guess I was wrong, since I
checked and the PBVH is clearly being rebuilt after the trim operation,
so it seems there isn't anything else weird going on here.
Rather than trying to figure that out, I'll just add the empty check
back here. Eventually we should completely replace the algorithm that
creates the PBVH nodes anyway.
