| Age | Commit message (Collapse) | Author |
|---|
|
pbvh->leaf_limit needs to be at least 4 to split nodes
original face boundaries properly.
|
|
BKE_sculpt_mask_layers_ensure now takes a depsgraph argument and
will evaluate the depsgraph if a multires mask layer is added.
This is necassary to update the multires runtime data so that
pbvh knows it has a grids mask layer.
Also added code to update pbvh->gridkey.
|
|
PBVH_FACES and PBVH_GRIDS do not store faces directly in nodes;
instead they store 'primitives', which are tesselation triangles
for PBVH_FACES and grids (which are per-loop) for PBVH_GRIDS.
Primitives from the same face could sometimes end up in different
PBVH nodes. This is now prevented in two ways:
* All primitives of the same face are given the same boundary
during PBVH build. This prevents them from being swapped
away from each other during partitioning.
* build_sub adjusts the final partition midpoint to fall
between primitives of different faces.
|
|
|
|
|
|
Cleaned up Dyntopo original triangle API (which is deprecated):
* BMVerts for original triangles are now stored.
* BKE_pbvh_bmesh_update_topology now handles original triangle
* data properly.
* BKE_pbvh_bmesh_node_save_orig can now initialize the original
coordinates from the current BMLogEntry.
* Ray casting of original data now returns active vertex.
Should fix various random crashes.
Hopefully this will fix a number of bugs.
|
|
Even if multires in sculpt mode doesn't yet support color
attributes, we should at least upload white color to avoid
making everything black.
Also fixed a bug where multires PBVHs didn't have access to
their CustomData attribute layout, which PBVH draw needs.
|
|
This is the conventional way of dealing with unused arguments in C++.
Also quiet enum conversion warnings.
|
|
Rewrite PBVH draw to allocate attributes into individual VBOs.
The old system tried to create a single VBO that could feed
every open viewport. This required uploading every color and
UV attribute to the viewport whether needed or not, often exceeding
the VBO limit.
This new system creates one VBO per attribute. Each attribute layout is
given its own GPU batch which is cached inside the owning PBVH node.
Notes:
* This is a full C++ rewrite. The old code is still there; ripping it out
can happen later.
* PBVH nodes now have a collection of batches, PBVHBatches, that keeps
track of all the batches inside the node.
* Batches are built exclusively from a list of attributes.
* Each attribute has its own VBO.
* Overlays, workbench and EEVEE can all have different attribute
layouts, each of which will get its own batch.
Reviewed by: Clement Foucault
Differential Revision: https://developer.blender.org/D15428
Ref D15428
|
|
To use function style cast '(unsigned char)x' can't be replaced by
'unsigned char(x)'.
|
|
|
|
Similar to the other refactors from T95965, this commit moves sculpt
face sets to use a generic integer attribute named `".sculpt_face_set"`.
This makes face sets accessible in the Python API.
The attribute is not visible in the attributes list or the spreadsheet
because it is meant for internal use, though that could be an option
in the future along with other similar attributes. Currently the change
is small, but in the future this could simplify code by allowing use
of more generic attribute APIs.
Differential Revision: https://developer.blender.org/D16045
|
|
New unified attribute API for sculpt code.
= Basic Design =
The sculpt attribute API can create temporary or permanent attributes (only supported in `PBVH_FACES` mode). Attributes are created via `BKE_sculpt_attribute_ensure.`
Attributes can be explicit CustomData attributes or simple array-based pseudo-attributes (this is useful for PBVH_GRIDS and PBVH_BMESH).
== `SculptAttributePointers` ==
There is a structure in `SculptSession` for convenience attribute pointers, `ss->attrs`. Standard attributes should assign these; the attribute API will automatically clear them when the associated attributes are released. For example, the automasking code stores its factor attribute layer in `ss->attrs.automasking_factor`.
== Naming ==
Temporary attributes should use the SCULPT_ATTRIBUTE_NAME macro for naming, it takes an entry in `SculptAttributePointers` and builds a layer name.
== `SculptAttribute` ==
Attributes are referenced by a special `SculptAttribute` structure, which holds
all the info needed to look up elements of an attribute at run time.
All of these structures live in a preallocated flat array in `SculptSession`, `ss->temp_attributes`. This is extremely important. Since any change to the `CustomData` layout can in principle invalidate every extant `SculptAttribute`, having them all in one block of memory whose location doesn't change allows us to update them transparently.
This makes for much simpler code and eliminates bugs. To see why this is tricky to get right, imagine we want to create three attributes in PBVH_BMESH mode and we provide our own `SculptAttribute` structs for the API to fill in. Each new layer will invalidate the `CustomData` block offsets in the prior one, leading to memory corruption.
Reviewed by: Brecht Van Lommel
Differential Revision: https://developer.blender.org/D15496
Ref D15496
|
|
Whether faces are hidden and face sets are orthogonal concepts, but
currently sculpt mode stores them together in the face set array.
This means that if anything is hidden, there must be face sets,
and if there are face sets, we have to keep track of what is hidden.
In other words, it adds a bunch of redundant work and state tracking.
On the user level it's nice that face sets and hiding are consistent,
but we don't need to store them together to accomplish that.
This commit uses the `".hide_poly"` attribute from rB2480b55f216c to
read and change hiding in sculpt mode. Face sets don't need to be
negative anymore, and a bunch of "face set <-> hide status" conversion
can be removed. Plus some other benefits:
- We don't need to allocate either array quite as much.
- The hide status can be read from 1/4 the memory as face sets.
- Updates when entering or exiting sculpt mode can be removed.
- More opportunities for early-outs when nothing is hidden.
- Separating concerns makes sculpt code more obvious.
- It will be easier to convert face sets into a generic int attribute.
Differential Revision: https://developer.blender.org/D15950
|
|
Caused by b5f7af31d6d4, which exposed the fact that the PBVH wasn't
retrieving the updated hide status attributes if they were allocated in
sculpt mode. Previously the attributes were always allocated when
entering sculpt mode.
|
|
corresponding data layers and using their values for computations.
Avoiding that should increase performance in many operations that
would otherwise have to read, write, or propagate these values.
It also means decreased memory usage-- not just for sculpt mode
but for any mesh that was in sculpt mode. Previously the mask, face set,
and hide status layers were *always* allocated by sculpt mode.
Here are a few basic tests when masking and face sets are not used:
| Test | Before | After |
| Subsurf Modifier | 148 ms | 126 ms |
| Sculpt Overlay Extraction | 24 ms every redraw | 0 ms |
| Memory usage | 252 MB | 236 MB |
I wouldn't expect any difference when they are used though.
The code changes are mostly just making sculpt features safe for when
the layers aren't stored, and some changes to the conversion to and
from the hide layers. Use of the ".hide_poly" attribute replaces testing
whether face sets are negative in many places.
Differential Revision: https://developer.blender.org/D15937
|
|
Follows existing naming for the most part, also use "num" as a suffix
in some instances (following our naming conventions).
|
|
This patch moves material indices from the mesh `MPoly` struct to a
generic integer attribute. The builtin material index was already
exposed in geometry nodes, but this makes it a "proper" attribute
accessible with Python and visible in the "Attributes" panel.
The goals of the refactor are code simplification and memory and
performance improvements, mainly because the attribute doesn't have
to be stored and processed if there are no materials. However, until
4.0, material indices will still be read and written in the old
format, meaning there may be a temporary increase in memory usage.
Further notes:
* Completely removing the `MPoly.mat_nr` after 4.0 may require
changes to DNA or introducing a new `MPoly` type.
* Geometry nodes regression tests didn't look at material indices,
so the change reveals a bug in the realize instances node that I fixed.
* Access to material indices from the RNA `MeshPolygon` type is slower
with this patch. The `material_index` attribute can be used instead.
* Cycles is changed to read from the attribute instead.
* BMesh isn't changed in this patch. Theoretically it could be though,
to save 2 bytes per face when less than two materials are used.
* Eventually we could use a 16 bit integer attribute type instead.
Ref T95967
Differential Revision: https://developer.blender.org/D15675
|
|
When allocating new `CustomData` layers, often we do redundant
initialization of arrays. For example, it's common that values are
allocated, set to their default value, and then set to some other
value. This is wasteful, and it negates the benefits of optimizations
to the allocator like D15082. There are two reasons for this. The
first is array-of-structs storage that makes it annoying to initialize
values manually, and the second is confusing options in the Custom Data
API. This patch addresses the latter.
The `CustomData` "alloc type" options are rearranged. Now, besides
the options that use existing layers, there are two remaining:
* `CD_SET_DEFAULT` sets the default value.
* Usually zeroes, but for colors this is white (how it was before).
* Should be used when you add the layer but don't set all values.
* `CD_CONSTRUCT` refers to the "default construct" C++ term.
* Only necessary or defined for non-trivial types like vertex groups.
* Doesn't do anything for trivial types like `int` or `float3`.
* Should be used every other time, when all values will be set.
The attribute API's `AttributeInit` types are updated as well.
To update code, replace `CD_CALLOC` with `CD_SET_DEFAULT` and
`CD_DEFAULT` with `CD_CONSTRUCT`. This doesn't cause any functional
changes yet. Follow-up commits will change to avoid initializing
new layers where the correctness is clear.
Differential Revision: https://developer.blender.org/D15617
|
|
The number of attribute domains isn't an attribute domain, so storing
ATTR_DOMAIN_NUM in a variable with an eAttrDomain type isn't correct.
In the cases it was used, the value wouldn't be accessed anyway.
|
|
|
|
* Fixed crash in debug draw code. Apparently this is
only used by PBVH draw?
* Debug draw code can now be forcibly enabled in release
mode (i.e. RelWithDebugInfo) by uncommenting a commented
out #define.
* Fixed colors in debug draw mode.
* PBVH node boxes in debug mode now flash a different color
when they are updated.
|
|
This commit moves the hide status of mesh vertices, edges, and faces
from the `ME_FLAG` to optional generic boolean attributes. Storing this
data as generic attributes can significantly simplify and improve code,
as described in T95965.
The attributes are called `.hide_vert`, `.hide_edge`, and `.hide_poly`,
using the attribute name semantics discussed in T97452. The `.` prefix
means they are "UI attributes", so they still contain original data
edited by users, but they aren't meant to be accessed procedurally by
the user in arbitrary situations. They are also be hidden in the
spreadsheet and the attribute list by default,
Until 4.0, the attributes are still written to and read from the mesh
in the old way, so neither forward nor backward compatibility are
affected. This means memory requirements will be increased by one byte
per element when the hide status is used. When the flags are removed
completely, requirements will decrease when hiding is unused.
Further notes:
* Some code can be further simplified to skip some processing when the
hide attributes don't exist.
* The data is still stored in flags for `BMesh`, necessitating some
complexity in the conversion to and from `Mesh`.
* Access to the "hide" property of mesh elements in RNA is slower.
The separate boolean arrays should be used where possible.
Ref T95965
Differential Revision: https://developer.blender.org/D14685
|
|
Instead of passing pointers to specific mesh data, rely on
retrieving that data from the mesh internally. This makes
it easier to support retrieving additional data from Mesh
(like active attribute names in D15101 or D15169). It also makes
the functions simpler conceptually, because they're drawing
a mesh with an acceleration strcture on top.
The BKE_id_attribute_copy_domains_temp call was unnecessary
because the GPU_pbvh_mesh_buffers_update function was only
called when Mesh/PBVH_FACES is used in the first place.
Differential Revision: https://developer.blender.org/D15197
|
|
|
|
Mask and color brushes were using the existing PBVH vertex "update tag"
to mark their modifications. This was mostly unnecessary, and causes
unnecessary calculation of normals. It also caused errors though,
because they didn't tag the corresponding PBVH node for normal
recalculation, causing problems on the borders of nodes, since one
node might accumulate into another's vertex normals, but the other
node wouldn't also accumulate and normalize the normals.
The solution is to only use the update tag for tagging deformed
vertices that need recalculated normals. Everything else is handled at
the PBVH node level (which was already the case, but it wasn't clear).
The update tag was also used for undo to tag the nodes corresponding to
changed vertices. This was wrong though, because normals and visibility
would also be recalculated for just color or mask undo steps. Instead,
just use local arrays to map from vertices to nodes.
Differential Revision: https://developer.blender.org/D15581
|
|
This is a port of sculpt-dev's `SculptVertRef` refactor
(note that `SculptVertRef was renamed to PBVHVertRef`)
to master. `PBVHVertRef` is a structure that abstracts
the concept of a vertex in the sculpt code; it's simply
an `intptr_t` wrapped in a struct.
For `PBVH_FACES` and `PBVH_GRIDS` this struct stores a
vertex index, but for `BMesh` it stores a direct pointer
to a BMVert. The intptr_t is wrapped in a struct to prevent
the accidental usage of it as an index.
There are many reasons to do this:
* Right now `BMesh` verts are not logical sculpt verts;
to use the sculpt API they must first be converted to indices.
This requires a lot of indirect lookups into tables, leading to performance
loss. It has also led to greater code complexity and duplication.
* Having an abstract vertex type makes it feasible to have one unified
temporary attribute API for all three PBVH modes, which in turn
made it rather trivial to port sculpt brushes to DynTopo in
sculpt-dev (e.g. the layer brush, draw sharp, the smooth brushes,
the paint brushes, etc). This attribute API will be in a future patch.
* We need to do this anyway for the eventual move to C++.
Differential Revision: https://developer.blender.org/D14272
Reviewed By: Brecht Van Lommel
Ref D14272
|
|
The recent multires winding fix missed a code branch.
|
|
|
|
Winding order of grid quads was backwards.
|
|
|
|
PBVH draw has an optimization where it only sends the
active attribute to the GPU in workbench mode. This
fails if multiple viewports are open with a mix of
workbench and EEVEE mode; it also causes severe lag
if any workbench viewport is in material mode.
There are two solutions: either add the code in sculpt-dev
that checks for EEVEE viewports at the beginning of each frame,
or integrate pbvh draw properly inside the draw manager
and let it handle which attributes should go to the GPU.
|
|
|
|
|
|
Moved gpu vert format checking outside of pbvh_update_draw_buffers,
which isn't called in every code path of BKE_pbvh_draw_cb. This led
to the draw cache being partially populated by old draw buffers
that were subsequently freed, causing a crash.
|
|
This patch adds support for PBVH drawing in EEVEE.
Notes:
# PBVH_FACES only. For Multires we'll need an API to get/cache attributes. DynTopo support will be merged in later with sculpt-dev's DynTopo implementation.
# Supports vertex color and UV attributes only; other types can be added fairly easily though.
# Workbench only sends the active vertex color and UV layers to the GPU.
# Added a new draw engine API method, DRW_cdlayer_attr_aliases_add. Please review.
# The vertex format object is now stored in the pbvh.
Reviewed By: Clément Foucault & Brecht Van Lommel & Jeroen Bakker
Differential Revision: https://developer.blender.org/D13897
Ref D13897
|
|
|
|
- CustomDataType -> eCustomDataType
- CustomDataMask -> eCustomDataMask
- AttributeDomain -> eAttrDomain
- NamedAttributeUsage -> eNamedAttrUsage
|
|
The final normalization step of sculpt normal calculation iterates over
all unique vertices in each node and marks them as done. However,
storing the done mask in a bitmap meant that multiple threads could
write to a single byte at the same time, because the bits for separate
threads could be stored in the same byte. This is not threadsafe
Fixing this issue seems to improve performance as well. First I tested
just clearing the entire bitmap after the normal calculation. Then I
tested using an array of booleans instead, which turned out to be
slightly better, and simplifies code a bit.
I tested on a Ryzen 3800x, on an 8 million polygon subdivided
Suzanne by using the grab brush with a radius large enough to
affect most of the mesh.
| Original | Clear Entire Bitmap | Boolean Array |
| --------- | ------------------- | ------------- |
| 67.9 ms | 59.1 ms | 57.9 ms |
| 1.0x | 1.15x | 1.17x |
Differential Revision: https://developer.blender.org/D14985
|
|
The PBVH draw cache wasn't being invalidated in
all cases. It is now invalidated whenever a PBVH
node's draw buffers are freed.
|
|
The "PROP" in the name reflects its generic status, and removing
"LOOP" makes sense because it is no longer associated with just
mesh face corners. In general the goal is to remove extra semantic
meaning from the custom data types.
|
|
This patch contains an initial pixel extractor for PBVH and an initial paint brush implementation.
PBVH is an accelleration structure blender uses internally to speed up 3d painting operations.
At this moment it is extensively used by sculpt, vertex painting and weight painting.
For the 3d texturing brush we will be using the PBVH for texture painting.
Currently PBVH is organized to work on geometry (vertices, polygons and triangles).
For texture painting this should be extended it to use pixels.
{F12995467}
Screen recording has been done on a Mac Mini with a 6 core 3.3 GHZ Intel processor.
# Scope
This patch only contains an extending uv seams to fix uv seams. This is not actually we want, but was easy to add
to make the brush usable.
Pixels are places in the PBVH_Leaf nodes. We want to introduce a special node for pixels, but that will be done
in a separate patch to keep the code review small. This reduces the painting performance when using
low and medium poly assets.
In workbench textures aren't forced to be shown. For now use Material/Rendered view.
# Rasterization process
The rasterization process will generate the pixel information for a leaf node. In the future those
leaf nodes will be split up into multiple leaf nodes to increase the performance when there
isn't enough geometry. For this patch this was left out of scope.
In order to do so every polygon should be uniquely assigned to a leaf node.
For each leaf node
for each polygon
If polygon not assigned
assign polygon to node.
Polygons are to complicated to be used directly we have to split the polygons into triangles.
For each leaf node
for each polygon
extract triangles from polygon.
The list of triangles can be stored inside the leaf node. The list of polygons aren't needed anymore.
Each triangle has:
poly_index.
vert_indices
delta barycentric coordinate between x steps.
Each triangle is rasterized in rows. Sequential pixels (in uv space) are stored in a single structure.
image position
barycentric coordinate of the first pixel
number of pixels
triangle index inside the leaf node.
During the performed experiments we used a fairly simple rasterization process by
finding the UV bounds of an triangle and calculate the barycentric coordinates per
pixel inside the bounds. Even for complex models and huge images this process is
normally finished within 0.5 second. It could be that we want to change this algorithm
to reduce hickups when nodes are initialized during a stroke.
Reviewed By: brecht
Maniphest Tasks: T96710
Differential Revision: https://developer.blender.org/D14504
|
|
color attribute system.
This commit removes sculpt colors from experimental
status and unifies it with vertex colors. It
introduces the concept of "color attributes", which
are any attributes that represents colors. Color
attributes can be represented with byte or floating-point
numbers and can be stored in either vertices or
face corners.
Color attributes share a common namespace
(so you can no longer have a floating-point
sculpt color attribute and a byte vertex color
attribute with the same name).
Note: this commit does not include vertex paint mode,
which is a separate patch, see:
https://developer.blender.org/D14179
Differential Revision: https://developer.blender.org/D12587
Ref D12587
|
|
Use a shorter/simpler license convention, stops the header taking so
much space.
Follow the SPDX license specification: https://spdx.org/licenses
- C/C++/objc/objc++
- Python
- Shell Scripts
- CMake, GNUmakefile
While most of the source tree has been included
- `./extern/` was left out.
- `./intern/cycles` & `./intern/atomic` are also excluded because they
use different header conventions.
doc/license/SPDX-license-identifiers.txt has been added to list SPDX all
used identifiers.
See P2788 for the script that automated these edits.
Reviewed By: brecht, mont29, sergey
Ref D14069
|
|
This is part of the project of converting `MVert` into `float3`.
(more details in T93602), The pbvh update flag is removed and
replaced with a bitmap stored in the PBVH structure. This
patch is similar to D13878. This is mainly setup for an eventual
performance improvement by removing the extra data from
mesh vertices, but if it's consistent with testing in the other patch
doing the same thing for another "temp tag", then it may actually
increase the speed of sculpt code slightly, since less memory needs
to be loaded when checking/changing the flags.
Differential Revision: https://developer.blender.org/D14000
|
|
Since d9c6ceb3b88b6db87490b08e0089f9a18e6c52d6 partial updates to
normals in sculpt-mode were accumulating into the current normal
instead of a zeroed value.
Zero vertex normal values tagged for calculation before accumulation.
Reviewed By: HooglyBoogly
Ref D13975
|
|
|
|
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
|
|
- Added space below non doc-string comments to make it clear
these aren't comments for the symbols directly below them.
- Use doxy sections for some headers.
- Minor improvements to doc-strings.
Ref T92709
|
|
|
