| Age | Commit message (Collapse) | Author |
|---|
|
|
|
As described in T91186, this commit moves mesh vertex normals into a
contiguous array of float vectors in a custom data layer, how face
normals are currently stored.
The main interface is documented in `BKE_mesh.h`. Vertex and face
normals are now calculated on-demand and cached, retrieved with an
"ensure" function. Since the logical state of a mesh is now "has
normals when necessary", they can be retrieved from a `const` mesh.
The goal is to use on-demand calculation for all derived data, but
leave room for eager calculation for performance purposes (modifier
evaluation is threaded, but viewport data generation is not).
**Benefits**
This moves us closer to a SoA approach rather than the current AoS
paradigm. Accessing a contiguous `float3` is much more efficient than
retrieving data from a larger struct. The memory requirements for
accessing only normals or vertex locations are smaller, and at the
cost of more memory usage for just normals, they now don't have to
be converted between float and short, which also simplifies code
In the future, the remaining items can be removed from `MVert`,
leaving only `float3`, which has similar benefits (see T93602).
Removing the combination of derived and original data makes it
conceptually simpler to only calculate normals when necessary.
This is especially important now that we have more opportunities
for temporary meshes in geometry nodes.
**Performance**
In addition to the theoretical future performance improvements by
making `MVert == float3`, I've done some basic performance testing
on this patch directly. The data is fairly rough, but it gives an idea
about where things stand generally.
- Mesh line primitive 4m Verts: 1.16x faster (36 -> 31 ms),
showing that accessing just `MVert` is now more efficient.
- Spring Splash Screen: 1.03-1.06 -> 1.06-1.11 FPS, a very slight
change that at least shows there is no regression.
- Sprite Fright Snail Smoosh: 3.30-3.40 -> 3.42-3.50 FPS, a small
but observable speedup.
- Set Position Node with Scaled Normal: 1.36x faster (53 -> 39 ms),
shows that using normals in geometry nodes is faster.
- Normal Calculation 1.6m Vert Cube: 1.19x faster (25 -> 21 ms),
shows that calculating normals is slightly faster now.
- File Size of 1.6m Vert Cube: 1.03x smaller (214.7 -> 208.4 MB),
Normals are not saved in files, which can help with large meshes.
As for memory usage, it may be slightly more in some cases, but
I didn't observe any difference in the production files I tested.
**Tests**
Some modifiers and cycles test results need to be updated with this
commit, for two reasons:
- The subdivision surface modifier is not responsible for calculating
normals anymore. In master, the modifier creates different normals
than the result of the `Mesh` normal calculation, so this is a bug
fix.
- There are small differences in the results of some modifiers that
use normals because they are not converted to and from `short`
anymore.
**Future improvements**
- Remove `ModifierTypeInfo::dependsOnNormals`. Code in each modifier
already retrieves normals if they are needed anyway.
- Copy normals as part of a better CoW system for attributes.
- Make more areas use lazy instead of eager normal calculation.
- Remove `BKE_mesh_normals_tag_dirty` in more places since that is
now the default state of a new mesh.
- Possibly apply a similar change to derived face corner normals.
Differential Revision: https://developer.blender.org/D12770
|
|
Ref T92709
|
|
|
|
|
|
Also hyphenate 'mouse-move' use doxy sections in render_update.c &
move function comment from the header to the source.
|
|
|
|
Use C comments for plain text.
|
|
Also use doxy style function reference `#` prefix chars when
referencing identifiers.
|
|
Also resolve a warning from the previous commit. The next blocker to
using const is `BKE_mesh_wrapper_ensure_mdata`.
|
|
|
|
|
|
It is more appropriate that `depths` is freed in `ED_view3d_depths_free`.
|
|
The depth cache (located in `RegionView3D::depths`) is used for quick
and simple occlusion testing in:
- particle selection,
- "Draw Curve" operator and
- "Interactive Light Track to Cursor" operator,
However, keeping a texture buffer in cache is not a recommended practice.
For displays with high resolution like 8k this represents something
around 132MB.
Also, currently, each call to `ED_view3d_depth_override` invalidates
the depth cache. So that depth is never reused in multiple calls from
an operator (this was not the case in blender 2.79).
This commit allows to create a depth cache and release it in the same
operator. Thus, the buffer is kept in cache for a short time, freeing
up space.
No functional changes.
|
|
When projecting into screen space Z value isn't always needed.
Add 2D projection functions, renaming them to avoid accidents
happening again.
- Add GPU_matrix_project_2fv
- Add ED_view3d_project_v2
- Rename ED_view3d_project to ED_view3d_project_v3
- Use the 2D versions of these functions when the Z value isn't used.
|
|
The check to include particle edit mode in the object-mode drop-down
didn't match the poll function to edit particle edit mode.
Share the check between both functions.
|
|
Now `ED_view3d_backbuf_depth_validate`, `ED_view3d_draw_depth` and
`ED_view3d_draw_depth_gpencil` are unified in `ED_view3d_depth_override`.
This new function replaces `ED_view3d_autodist_init`.
Also, since `ED_view3d_depth_update` depends on the render context, and
changing the context is a slow operation, that function also was removed,
and the depth buffer cached is now updated inside the new unified drawing
function when the "bool update_cache" parameter is true.
Finally `V3D_INVALID_BACKBUF` flag has been renamed and moved to
`runtime.flag`.
Differential revision: https://developer.blender.org/D10678
|
|
|
|
|
|
Undo/redo was crashing & asserting.
There ware two bugs:
- Entering particle system edit-mode requires an evaluated depsgraph.
- The active object could get out of sync when undo/redo
moves between different modes.
The new test 'view3d_mesh_particle_edit_mode_simple'
in `tests/ui_simulate/test_undo.py` exposes both issues.
|
|
|
|
|
|
In two cases the percentage property was actually used incorrectly, as
pointed out in T82070. The range was [0, 1], but the properties were still
displayed as percentages.
There is a preference to control whether to display factors or
percentages, so it usually doesn't make sense to manually define
properties as percentages.
Resolves T82070
Differential Revision: https://developer.blender.org/D9344
|
|
Comment blocks not conforming to convention.
|
|
The hair mirroring code seems to expect that hair is emitted from faces.
The PE_mirror_x contains the following expression: mirrorfaces[pa->num * 2].
This only makes sense when pa->num is a face index.
The simplest short term solution is to disable the mirror operator when
the particles haven't been emitted from faces.
Diffferential Revision: https://developer.blender.org/D10002
|
|
|
|
The issue was that the same point cache was read by multiple
threads at the same time (the same object was evaluated for
render and for the viewport).
Both threads incremented PTCacheMem->cur which lead to the crash.
The fix is to remove the PTCacheMem->cur and store it on the
stack instead. This way every thread has its own cur.
Reviewers: brecht
Differential Revision: https://developer.blender.org/D9606
|
|
|
|
Corrects incorrect usages of the words 'then' and 'than'.
Differential Revision: https://developer.blender.org/D9246
Reviewed by Campbell Barton
|
|
Follow our code style guide by using C-comments for text descriptions.
|
|
Deleting the particle editmode is fully supported by undo now.
No need to warn the user about that anymore.
|
|
We have our own assert implementation, `BLI_assert()` that is prefered over the
C standard library one. Its output is more consistent across compilers and
makes termination on assert failure optional (through `WITH_ASSERT_ABORT`).
In many places we'd include the C library header without ever accessing it.
|
|
Cleanup old tracker task format to the new. e.g: [#34039] to T34039
Ref D8718
|
|
This adds XYZ symmetry as a property of meshes and updates all modes to
use the mesh symmetry by default to have a consistent tool behavior
between all modes and when switching objects.
Reviewed By: brecht, mano-wii, campbellbarton
Maniphest Tasks: T79785
Differential Revision: https://developer.blender.org/D8587
|
|
situations
In the situation that the PTCacheEdit is not available (e.g. when
editing unbaked Cloth or Softbody caches - or, as in the report, being
in Cloth or Softbody Editing Type for Hair particles), accesing the
corresponding ParticleSystem or ParticleSystemModifierData would crash.
Now access these later (after PE_start_edit had the chance to return
early on this non-valid situation).
Reviewers: JacquesLucke
Maniphest Tasks: T80900
Differential Revision: https://developer.blender.org/D8941
|
|
Remove redundantly nested `#if` and `#ifdef` statements.
One nested `#if 0` block was left untouched, as it's in particle code
that's no longer maintained. Furthermore, that block also has some
explanation as to the differences between the enabled & disabled parts.
One nested `#if 0` construct was completely removed, leaving only the
actually used bit of code. There was no explanation as to the usefulness
of the disabled code, and it hasn't been touched in years.
No functional changes.
|
|
We now use GPU_blend for enabling / disabling blending and explicitly
set the blend equation.
|
|
Also order sizeof(..) first to promote other values to size_t.
|
|
The abbreviation 'init' is brief, unambiguous and already used
in thousands of places, also initialize is often accidentally
written with British spelling.
|
|
This addresses warnings from Clang-Tidy's `readability-else-after-return`
rule in the `source/blender/editors` module.
No functional changes.
|
|
Caused by short overflow, change to int instead.
Maniphest Tasks: T78415
Differential Revision: https://developer.blender.org/D8148
|
|
|
|
This caused an additional argument when exiting object modes
and many other low level functions which don't need to access context.
This simplifies fixing T77073.
|
|
|
|
|
|
Similar to T76859
|
|
|
|
|
|
There was no reason to use a short here,
this was just a convention from existing code.
|
|
- 'coords' is an abbreviation for coordinates, not 'cords'.
- Rename 'moves' to 'coords_len'.
|
