The issue was uncovered by the 0f89bcdbeb, but the root cause goes
into a much earlier design violation happened in the code: the modifier
evaluation function is modifying input mesh, which is not something
what is ever expected.
Bring code closer to the older state where such modification is only
done for the object in edit mode.
---
From own tests works seems to work fine, but extra eyes and testing
is needed.
Differential Revision: https://developer.blender.org/D14191
New code from the vertex normal refactor cfa53e0fbe combined with older code
from 592759e3d6 that disabled instancing for custom normals and autosmooth
meant that instancing was always disabled.
However we do not need to disable instancing for custom normals and autosmooth
at all, this can be shared between instances just fine.
It's better not to expose the details of where the dirty flags are
stored to every place that wants to know if the normals are dirty.
Some of these places are relics from before vertex normals were
computed lazily anyway, so this is more of an incrememtal cleanup.
This will make part of the fix for T95839 simpler.
Currently, whenever any BMesh is converted to a Mesh (except for edit
mode switching), original index (`CD_ORIGINDEX`) layers are added.
This is incorrect, because many operations just convert some Mesh into
a BMesh and then back, but they shouldn't make any assumption about
where their input mesh came from. It might even come from a primitive
in geometry nodes, where there are no original indices at all.
Conceptually, mesh original indices should be filled by the modifier
stack when first creating the evaluated mesh. So that's where they're
moved in this patch. A separate function now fills the indices with their
default (0,1,2,3...) values. The way the mesh wrapper system defers
the BMesh to Mesh conversion makes this a bit less obvious though.
The old behavior is incorrect, but it's also slower, because three
arrays the size of the mesh's vertices, edges, and faces had to be
allocated and filled during the BMesh to Mesh conversion, which just
ends up putting more pressure on the cache. In the many cases where
original indices aren't used, I measured an **8% speedup** for the
conversion (from 76.5ms to 70.7ms).
Generally there is an assumption that BMesh is "original" and Mesh is
"evaluated". After this patch, that assumption isn't quite as strong,
but it still exists for two reasons. First, original indices are added
whenever converting a BMesh "wrapper" to a Mesh. Second, original
indices are not added to the BMesh at the beginning of evaluation,
which assumes that every BMesh in the viewport is original and doesn't
need the mapping.
Differential Revision: https://developer.blender.org/D14018
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 partially caused by a stupid mistake in cfa53e0fbe
where I missed initializing the `vert_normals` pointer in
`MResolvePixelData`. It's also caused by questionable assumptions
from DerivedMesh code that vertex normals would be valid.
The fix used here is to create a temporary mesh with the data necessary
to compute vertex normals, and ensure them here. This is used because
normal calculation is only implemented for `Mesh` and edit mesh, not
`DerivedMesh`. While this might not be great for performance, it's
potentially aligned with future refactoring of this code to remove
`DerivedMesh` completely. Since this is one of the last places the data
structure is used, that would be a great improvement.
Differential Revision: https://developer.blender.org/D13960
Since we have a node that sets a mesh's auto smooth angle
(unfortunately, in retrospect), we generally can't assume at all
that value is the same as whatever input mesh. Similar asserts
were removed previously in 8216b759e9. While the attempt
at assertions to clarify assumptions is noble, this one doesn't
make sense anymore.
I found this while investigating T95479.
Differential Revision: https://developer.blender.org/D14009
Since we have a node that sets a mesh's auto smooth angle
(unfortunately, in retrospect), we generally can't assume at all
that value is the same as whatever input mesh. Similar asserts
were removed previously in 8216b759e9. While the attempt
at assertions to clarify assumptions is noble, this one doesn't
make sense anymore.
I found this while investigating T95479.
Differential Revision: https://developer.blender.org/D14009
This is partially caused by a stupid mistake in cfa53e0fbe
where I missed initializing the `vert_normals` pointer in
`MResolvePixelData`. It's also caused by questionable assumptions
from DerivedMesh code that vertex normals would be valid.
The fix used here is to create a temporary mesh with the data necessary
to compute vertex normals, and ensure them here. This is used because
normal calculation is only implemented for `Mesh` and edit mesh, not
`DerivedMesh`. While this might not be great for performance, it's
potentially aligned with future refactoring of this code to remove
`DerivedMesh` completely. Since this is one of the last places the data
structure is used, that would be a great improvement.
Differential Revision: https://developer.blender.org/D13960
Remove functions and function pointers that were never set or never
used at all. The "tessface" original index handling in `subsurf_ccg.c`
can be removed because the data was never used.
This function and flags weren't used outside of DerivedMesh
code, and since the plan is to remove the data structure, it makes
sense to remove complexity where possible.
The evaluated mesh is a result of evaluated modifiers, and referencing
other evaluated IDs such as materials.
It can not be stored in the EditMesh structure which is intended to be
re-used by many areas. Such sharing was causing ownership errors causing
bugs like
T93855: Cycles crash with edit mode and simultaneous viewport and final render
The proposed solution is to store the evaluated edit mesh and its cage in
the object's runtime field. The motivation goes as following:
- It allows to avoid ownership problems like the ones in the linked report.
- Object level is chosen over mesh level is because the evaluated mesh
is affected by modifiers, which are on the object level.
This patch allows to have modifier stack of an object which shares mesh with
an object which is in edit mode to be properly taken into account (before
the change the modifier stack from the active object will be used for all
objects which share the mesh).
There is a change in the way how copy-on-write is handled in the edit mode to
allow proper state update when changing active scene (or having two windows
with different scenes). Previously, the copt-on-write would have been ignored
by skipping tagging CoW component. Now it is ignored from within the CoW
operation callback. This allows to update edit pointers for objects which are
not from the current depsgraph and where the edit_mesh was never assigned in
the case when the depsgraph was evaluated prior the active depsgraph.
There is no user level changes changes expected with the CoW handling changes:
should not affect on neither performance, nor memory consumption.
Tested scenarios:
- Various modifiers configurations of objects sharing mesh and be part of the
same scene.
- Steps from the reports: T93855, T82952, T77359
This also fixes T76609, T72733 and perhaps other reports.
Differential Revision: https://developer.blender.org/D13824
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
This patch implements the vector types (i.e:`float2`) by making heavy
usage of templating. All vector functions are now outside of the vector
classes (inside the `blender::math` namespace) and are not vector size
dependent for the most part.
In the ongoing effort to make shaders less GL centric, we are aiming
to share more code between GLSL and C++ to avoid code duplication.
####Motivations:
- We are aiming to share UBO and SSBO structures between GLSL and C++.
This means we will use many of the existing vector types and others
we currently don't have (uintX, intX). All these variations were
asking for many more code duplication.
- Deduplicate existing code which is duplicated for each vector size.
- We also want to share small functions. Which means that vector
functions should be static and not in the class namespace.
- Reduce friction to use these types in new projects due to their
incompleteness.
- The current state of the `BLI_(float|double|mpq)(2|3|4).hh` is a
bit of a let down. Most clases are incomplete, out of sync with each
others with different codestyles, and some functions that should be
static are not (i.e: `float3::reflect()`).
####Upsides:
- Still support `.x, .y, .z, .w` for readability.
- Compact, readable and easilly extendable.
- All of the vector functions are available for all the vectors types
and can be restricted to certain types. Also template specialization
let us define exception for special class (like mpq).
- With optimization ON, the compiler unroll the loops and performance
is the same.
####Downsides:
- Might impact debugability. Though I would arge that the bugs are
rarelly caused by the vector class itself (since the operations are
quite trivial) but by the type conversions.
- Might impact compile time. I did not saw a significant impact since
the usage is not really widespread.
- Functions needs to be rewritten to support arbitrary vector length.
For instance, one can't call `len_squared_v3v3` in
`math::length_squared()` and call it a day.
- Type cast does not work with the template version of the `math::`
vector functions. Meaning you need to manually cast `float *` and
`(float *)[3]` to `float3` for the function calls.
i.e: `math::distance_squared(float3(nearest.co), positions[i]);`
- Some parts might loose in readability:
`float3::dot(v1.normalized(), v2.normalized())`
becoming
`math::dot(math::normalize(v1), math::normalize(v2))`
But I propose, when appropriate, to use
`using namespace blender::math;` on function local or file scope to
increase readability.
`dot(normalize(v1), normalize(v2))`
####Consideration:
- Include back `.length()` method. It is quite handy and is more C++
oriented.
- I considered the GLM library as a candidate for replacement. It felt
like too much for what we need and would be difficult to extend / modify
to our needs.
- I used Macros to reduce code in operators declaration and potential
copy paste bugs. This could reduce debugability and could be reverted.
- This touches `delaunay_2d.cc` and the intersection code. I would like
to know @howardt opinion on the matter.
- The `noexcept` on the copy constructor of `mpq(2|3)` is being removed.
But according to @JacquesLucke it is not a real problem for now.
I would like to give a huge thanks to @JacquesLucke who helped during this
and pushed me to reduce the duplication further.
Reviewed By: brecht, sergey, JacquesLucke
Differential Revision: https://developer.blender.org/D13791
This patch implements the vector types (i.e:`float2`) by making heavy
usage of templating. All vector functions are now outside of the vector
classes (inside the `blender::math` namespace) and are not vector size
dependent for the most part.
In the ongoing effort to make shaders less GL centric, we are aiming
to share more code between GLSL and C++ to avoid code duplication.
####Motivations:
- We are aiming to share UBO and SSBO structures between GLSL and C++.
This means we will use many of the existing vector types and others
we currently don't have (uintX, intX). All these variations were
asking for many more code duplication.
- Deduplicate existing code which is duplicated for each vector size.
- We also want to share small functions. Which means that vector
functions should be static and not in the class namespace.
- Reduce friction to use these types in new projects due to their
incompleteness.
- The current state of the `BLI_(float|double|mpq)(2|3|4).hh` is a
bit of a let down. Most clases are incomplete, out of sync with each
others with different codestyles, and some functions that should be
static are not (i.e: `float3::reflect()`).
####Upsides:
- Still support `.x, .y, .z, .w` for readability.
- Compact, readable and easilly extendable.
- All of the vector functions are available for all the vectors types
and can be restricted to certain types. Also template specialization
let us define exception for special class (like mpq).
- With optimization ON, the compiler unroll the loops and performance
is the same.
####Downsides:
- Might impact debugability. Though I would arge that the bugs are
rarelly caused by the vector class itself (since the operations are
quite trivial) but by the type conversions.
- Might impact compile time. I did not saw a significant impact since
the usage is not really widespread.
- Functions needs to be rewritten to support arbitrary vector length.
For instance, one can't call `len_squared_v3v3` in
`math::length_squared()` and call it a day.
- Type cast does not work with the template version of the `math::`
vector functions. Meaning you need to manually cast `float *` and
`(float *)[3]` to `float3` for the function calls.
i.e: `math::distance_squared(float3(nearest.co), positions[i]);`
- Some parts might loose in readability:
`float3::dot(v1.normalized(), v2.normalized())`
becoming
`math::dot(math::normalize(v1), math::normalize(v2))`
But I propose, when appropriate, to use
`using namespace blender::math;` on function local or file scope to
increase readability.
`dot(normalize(v1), normalize(v2))`
####Consideration:
- Include back `.length()` method. It is quite handy and is more C++
oriented.
- I considered the GLM library as a candidate for replacement. It felt
like too much for what we need and would be difficult to extend / modify
to our needs.
- I used Macros to reduce code in operators declaration and potential
copy paste bugs. This could reduce debugability and could be reverted.
- This touches `delaunay_2d.cc` and the intersection code. I would like
to know @howardt opinion on the matter.
- The `noexcept` on the copy constructor of `mpq(2|3)` is being removed.
But according to @JacquesLucke it is not a real problem for now.
I would like to give a huge thanks to @JacquesLucke who helped during this
and pushed me to reduce the duplication further.
Reviewed By: brecht, sergey, JacquesLucke
Differential Revision: https://developer.blender.org/D13791
This patch implements the vector types (i.e:float2) by making heavy
usage of templating. All vector functions are now outside of the vector
classes (inside the blender::math namespace) and are not vector size
dependent for the most part.
In the ongoing effort to make shaders less GL centric, we are aiming
to share more code between GLSL and C++ to avoid code duplication.
Motivations:
- We are aiming to share UBO and SSBO structures between GLSL and C++.
This means we will use many of the existing vector types and others we
currently don't have (uintX, intX). All these variations were asking
for many more code duplication.
- Deduplicate existing code which is duplicated for each vector size.
- We also want to share small functions. Which means that vector functions
should be static and not in the class namespace.
- Reduce friction to use these types in new projects due to their
incompleteness.
- The current state of the BLI_(float|double|mpq)(2|3|4).hh is a bit of a
let down. Most clases are incomplete, out of sync with each others with
different codestyles, and some functions that should be static are not
(i.e: float3::reflect()).
Upsides:
- Still support .x, .y, .z, .w for readability.
- Compact, readable and easilly extendable.
- All of the vector functions are available for all the vectors types and
can be restricted to certain types. Also template specialization let us
define exception for special class (like mpq).
- With optimization ON, the compiler unroll the loops and performance is
the same.
Downsides:
- Might impact debugability. Though I would arge that the bugs are rarelly
caused by the vector class itself (since the operations are quite trivial)
but by the type conversions.
- Might impact compile time. I did not saw a significant impact since the
usage is not really widespread.
- Functions needs to be rewritten to support arbitrary vector length. For
instance, one can't call len_squared_v3v3 in math::length_squared() and
call it a day.
- Type cast does not work with the template version of the math:: vector
functions. Meaning you need to manually cast float * and (float *)[3] to
float3 for the function calls.
i.e: math::distance_squared(float3(nearest.co), positions[i]);
- Some parts might loose in readability:
float3::dot(v1.normalized(), v2.normalized())
becoming
math::dot(math::normalize(v1), math::normalize(v2))
But I propose, when appropriate, to use
using namespace blender::math; on function local or file scope to
increase readability. dot(normalize(v1), normalize(v2))
Consideration:
- Include back .length() method. It is quite handy and is more C++
oriented.
- I considered the GLM library as a candidate for replacement.
It felt like too much for what we need and would be difficult to
extend / modify to our needs.
- I used Macros to reduce code in operators declaration and potential
copy paste bugs. This could reduce debugability and could be reverted.
- This touches delaunay_2d.cc and the intersection code. I would like to
know @Howard Trickey (howardt) opinion on the matter.
- The noexcept on the copy constructor of mpq(2|3) is being removed.
But according to @Jacques Lucke (JacquesLucke) it is not a real problem
for now.
I would like to give a huge thanks to @Jacques Lucke (JacquesLucke) who
helped during this and pushed me to reduce the duplication further.
Reviewed By: brecht, sergey, JacquesLucke
Differential Revision: http://developer.blender.org/D13791
- 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
Evaluated meshes from curves are presented to render engines as
separate instance objects now, just like evaluated meshes from other
object types like point clouds and volumes. For that reason, cycles
should not consider curve objects as geometry (previously it did,
meaning it retrieved a second mesh from the curve object as well
as the temporary evaluated mesh geometry).
Further, avoid adding a curve object's evaluated mesh as data_eval,
since that is special behavior for meshes that is arbitrary. Adding an
evaluated mesh there but not an evalauted pointcloud is arbitrary,
for example. Retrieve the evaluated mesh in from the geometry set
in BKE_object_get_evaluated_mesh now, to support that change.
This gets us closer to a place where all of an object's evaluated data
is stored in geometry_set_eval, and we just have helper functions
to access specific geometry components.
Differential Revision: https://developer.blender.org/D13118
This commit removes the implicit conversion from points to a mesh
that used to happen before the next modifier. It also removes the
implicit realizing of instances that happened before another modifier.
Now we have specific nodes for both of these operations, the
implicit conversions make less sense, and implicit instance
realizing has already been removed in other nodes.
This adds another geometry nodes modifier before modifiers that would
have realized instances implicitly before. Currently adding another
data-block during versioning after linking means that an assert needs
to be changed. That should be made unnecessary by T92333.
Differential Revision: https://developer.blender.org/D12722
Remove the 'only_face_normals' argument.
- BKE_mesh_calc_normals_poly for polygon normals.
- BKE_mesh_calc_normals_poly_and_vertex for poly and vertex normals.
Order arguments logically:
- Pair array and length arguments.
- Position normal array arguments (to be filled) last.
This commit moves the storage of `bDeformGroup` and the active index
to `Mesh`, `Lattice`, and `bGPdata` instead of `Object`. Utility
functions are added to allow easy access to the vertex groups given
an object or an ID.
As explained in T88951, the list of vertex group names is currently
stored separately per object, even though vertex group data is stored
on the geometry. This tends to complicate code and cause bugs,
especially as geometry is created procedurally and tied less closely
to an object.
The "Copy Vertex Groups to Linked" operator is removed, since they
are stored on the geometry anyway.
This patch leaves the object-level python API for vertex groups in
place. Creating a geometry-level RNA API can be a separate step;
the changes in this commit are invasive enough as it is.
Note that opening a file saved in 3.0 in an earlier version means
the vertex groups will not be available.
Differential Revision: https://developer.blender.org/D11689
Rename:
- EDBM_mesh_free -> EDBM_mesh_free_data
BKE_editmesh_free -> BKE_editmesh_free_data
Since this doesn't free the edit-mesh pointer.
- BKE_editmesh_free_derivedmesh -> BKE_editmesh_free_derived_caches
Since this no longer uses derived-mesh, match naming for the related
object function BKE_object_free_derived_caches.
Also remove `do_tessellate` argument from BKE_editmesh_create,
since the caller can explicitly do this if it's needed,
with the advantage that it can be combined with normal calculation
which is faster on high-poly meshes.
This patch fixes many minor spelling mistakes, all in comments or
console output. Mostly contractions like can't, won't, don't, its/it's,
etc.
Differential Revision: https://developer.blender.org/D11663
Reviewed by Harley Acheson
- BKE_mesh_copy_parameters_for_eval to be used for evaluated meshes only
as it doesn't handle ID user-counts.
- BKE_mesh_copy_parameters is a general function for copying parameters
between meshes.
This patch adds initial curve support to geometry nodes. Currently
there is only one node available, the "Curve to Mesh" node, T87428.
However, the aim of the changes here is larger than just supporting
curve data in nodes-- it also uses the opportunity to add better spline
data structures, intended to replace the existing curve evaluation code.
The curve code in Blender is quite old, and it's generally regarded as
some of the messiest, hardest-to-understand code as well. The classes
in `BKE_spline.hh` aim to be faster, more extensible, and much more
easily understandable. Further explanation can be found in comments in
that file.
Initial builtin spline attributes are supported-- reading and writing
from the `cyclic` and `resolution` attributes works with any of the
attribute nodes. Also, only Z-up normal calculation is implemented
at the moment, and tilts do not apply yet.
**Limitations**
- For now, you must bring curves into the node tree with an "Object
Info" node. Changes to the curve modifier stack will come later.
- Converting to a mesh is necessary to visualize the curve data.
Further progress can be tracked in: T87245
Higher level design document: https://wiki.blender.org/wiki/Modules/Physics_Nodes/Projects/EverythingNodes/CurveNodes
Differential Revision: https://developer.blender.org/D11091
