Currently for node tools we create and evaluate a temporary depsgraph
with all the selected object data-blocks and all data-blocks referenced
by the node tree.
Needless to say, this can be very slow when those data-blocks contain
arbitrary procedural operations. Re-evaluating all the selected objects
is particularly because it will give a slowdown even in very basic uses
of node tools.
Originally I hoped that geometry nodes could be made to work with
original as well as evaluated data-blocks. But that would require far
too many tricky changes and arguably isn't right design-wise anyway.
Instead of that, this commit makes node tools dependency graph
evaluation more fine-grained in a few ways.
1. Remove the evaluation of selected objects. These are always visible
in the viewport and part of the active depsgraph anyway. To protect
against cyclic dependencies, we now compare `orig_id` instead of the
object pointer itself.
2. Evaluate the node group and its dependencies in a separate depsgraph
used only when necessary. This allows using the original node tree
without any copies when it doesn't reference any data-blocks.
3. Evaluate IDs from node group inputs (from the redo panel) in the extra
depsgraph as well, only when necessary.
Pull Request: https://projects.blender.org/blender/blender/pulls/120723
This makes the read and write API functions match more closely, and adds
asserts to check that the data size is as expected.
There are still a few places remaining that use BLO_read_data_address
and similar generic functions, these should eventually be replaced as well.
Pull Request: https://projects.blender.org/blender/blender/pulls/120994
Support for building blender with clang on windows on x64 was added
years ago but given there are no active users support has crumbled a
bit.
This PR brings the build system back into working order but upstream
patches in openVDB are still required for a successful build see PR
#120317 for details.
Blender when build with clang the classroom scenes rendered on the cpu
with cycles is seeing a 5% reduction in render time on both an
AMD 7700x and an Intel 14900k.
With the matrix socket being introduced into geometry nodes, we
are starting to deal with more complex transforms like perspective
projection. For those matrices projecting a point is not as simple
as just matrix multiplication, there has to be an additional normalization
step after. To solve that in an intuitive way consistent with how it's
typically solved in code, add a new "Project Point" node.
The canonical use case for now is in combination with the mouse
position, viewport transform, and raycast nodes, to find where the
mouse clicked on the edited geometry.
Pull Request: https://projects.blender.org/blender/blender/pulls/120597
Add a simple node to compute the intersection, difference, or union
between SDF grids. This should be the first new use case for the
new volume grid nodes that wasn't possible before.
For naming and multi-inputs, the node uses the same design as the
mesh boolean node. We considered splitting each operation into a
separate node, but though most users considered these different
"modes" of the same operation.
One thing to keep in mind is that it's important for the grids to
have exactly the same transform. If they have different transforms,
the second grid must be resampled to match the first, because the
OpenVDB CSG tools have that requirement. Resampling is expensive
(for SDF grids it means a grid -> mesh -> grid round trip) and should
be avoided.
Pull Request: https://projects.blender.org/blender/blender/pulls/118879
`NOD_zone_socket_items.hh` contained code for different nodes. It's better to
split this into headers per node, because that scales better. Also it helps to
keep the code for each individual node more closely together.
Pull Request: https://projects.blender.org/blender/blender/pulls/120945
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.
Improve working with multiple group input nodes by automatically
hiding new node group input sockets on any group input node with
hidden extension socket.
Whenever the extension socket is hidden, the user can now expect
the socket visibility of the group input node to stay fixed.
When the extension socket is shown on the other hand, the group input
node will update when sockets are added to the group interface.
Pull Request: https://projects.blender.org/blender/blender/pulls/120067
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
The compositor execute functions have a `rendering` argument to specify
if the compositor is executing as part of the render pipeline. But the
render context argument is null if we are not rendering, so the
`rendering` arguement is redundant and can be removed.
Additionally, we no longer use use_file_output as a hack to detect
rendering.
Pull Request: https://projects.blender.org/blender/blender/pulls/120659
Add utility function to handle supported types explicitly.
Right now Blur Attribute node generates all the attribute types except for bool.
So, there is redundant instances of functions for types:
float2, int2, int8_t, ColorGeometry4b, math::Quaternion, float4x4.
These types are not supported and should not be used for instantiation.
This makes the blender binary 42 KB smaller.
Pull Request: https://projects.blender.org/blender/blender/pulls/120639
Rewrite of Scale Elements. Main changes related with removing unnecessary
abstractions (like structures of fields). Next, by using grouping approach,
all data is represented as spans. This provide ability to unify code for
different domains. Using of general utils like IndexMask, Group processing
and array utils provides much more parallelism and better memory usage.
In result, this refactoring result in 4-10 average speed improvement in
attached benchmark file with different probability and scale of elements.
See the PR for more before/after timing information and a benchmark file.
Pull Request: https://projects.blender.org/blender/blender/pulls/115142
- "Log Encoding with Chroma inset and rotation": add "of primaries" in
the description of the AgX Log color space to better explain the
operation, based on wording in !106355.
- Remove a few double spaces.
- Make Line Art title case everywhere, to convey it's the system /
brand / product name and not the generic concept.
- "Copy Absolute coordinates or Normal vector" -> "of Normal Vector":
typo.
- "Makes a link between selected output in input sockets" ->
"Make...", "output and input": typo.
- "Purge Unused Data From This File" -> "from this": title case as per
HIGs.
- GPencil -> Grease Pencil: no reason to use an abbreviation here.
- "Around Current Frame" -> "Around Frame": actual name of the
onion-skinning method.
- "... (layer height for layer tool, i.e.)" -> "(i.e. the layer height
for the layer tool)": put "i.e." at the start of the sentence.
- Expand description of toe-in stereo camera option.
- "Children collections their parent-collection-specific settings" ->
"Children collections with their...": typo.
- "Generate vertex weights base on..." -> "based on" : typo, lower
case.
- Expand description of GP modifier properties, based on their mesh
counterparts.
- "AEnvelope" -> "Envelope": typo.
- "Falloff type the feather" -> "of the feather": typo.
- "usually make transition as long as effect strip": rephrase.
- "When disabled a users extensions directory is created" -> "a
user's": typo.
- "successfull" -> "successful": typo.
- "Remove all attributes... a single wildcard (*).": remove trailing
".".
- "..., use "Save Preferences."": remove trailing ".".
Some issues reported by Marina Veselkova and Tamar Mebonia.
Pull Request: https://projects.blender.org/blender/blender/pulls/120649
Previously, this conversion would often result in invalid quaternions or
hit an assert in `normalized_to_quat_fast`. It's not super nice to convert
to euler as an intermediate step performance wise, but it seems to be
the easiest solution for now. Extracting rotations from matrices should
not be done all that often anyway.
Pull Request: https://projects.blender.org/blender/blender/pulls/120568
Extract
- Statuses for the external text editor
- Newly created enum node item
- Newly created plane track data
- Newly created custom orientation data
- Operator names in drag and drop menu (need to use operator's
translation context)
- GN attribute statistic node inputs
Disambiguate
- Single-letter colors: A and B can mean Alpha and Blue, or simply A
and B as in two operands in an operation
- Dissolve: issue reported by Tamar Mebonia in #43295
- Translate in the User Preferences. This introduces a new
BLT_I18NCONTEXT_EDITOR_PREFERENCES ("Preferences") translation
context
- Planar (reported by deathblood)
This one is incomplete, because there is currently no way to
disambiguate presets or GN fields. I don't see how either could be
achieved cleanly.
The former would need to define the context inside the preset and
evaluate the file prior to showing it in the presets menu, which
sound bad.
The latter would need to introduce an additional string inside
`FieldInput`s, which would be controversial given how little it
would be used.
Remove
- Unused translation `iface_("%s")` in toolbar
- Remove obsolete N_() tags in a few node descriptions.
Pull Request: https://projects.blender.org/blender/blender/pulls/119065
To ease the process of debugging a node group while creating
a node tool, while the group is visible in a node editor, log the socket
values from its last execution. The values are only logged for the
active object is nothing selected. The Viewer node is still not supported
because visualization would probably be very tricky.
Pull Request: https://projects.blender.org/blender/blender/pulls/120596
Regular topology mapping does not care about deduplication of relations. And
this is correct for a lot of the case where we expect creating difference of this
relations in next operations. But this is not the case for users, since this should
be actually trivial info. Due to this reason, construct full topology map and count
number of unique connections. Performance side: difference linear check and
hash set for different sizes of connections set.
Pull Request: https://projects.blender.org/blender/blender/pulls/118848
Not all matrices are invertable (e.g. when a transformation matrix is has
a scale of zero), and it's sometimes important when a matrix is.
This also better specifies the behavior when the matrix is not invertable:
the identity matrix is returned.
Pull Request: https://projects.blender.org/blender/blender/pulls/120530
`interpolate_curve_attributes()` did not transfer `cyclic` attribute to
target curves. This fixes the the problem and allows Interpolate
Curves node to interpolate cyclic curves and also work between a cyclic
curve and a non-cyclic curve.
Pull Request: https://projects.blender.org/blender/blender/pulls/120460
To know if link is connected to dangling reroute and can be skipped
as value-less, we need to know if reroute is dangling. This requires
graph traversal. Currently this is done by non-recursive iteration.
But this can lead quadratic complexity for some of the cases.
Other way is to make this linear while cache building.
Pull Request: https://projects.blender.org/blender/blender/pulls/120375
This patch ports the GPU Vector Blur node to the CPU, which is in turn
ported from EEVEE. This is a breaking change since it produces different
motion blur results that are more similar to EEVEE's motion blur.
Further, the Curved, Minimum, and Maximum options were removed on the
user level since they are not used in the new implementation.
There are no significant changes to the code, except in the max velocity
computation as well as the velocity dilation passes. The GPU code uses
atomic indirection buffers, while the CPU runs single threaded for the
dilation pass, since it is a fast pass anyways. However, we impose
artificial constraints on the precision of the dilation process for
compatibility with the atomic implementation.
There are still tiny differences between CPU and GPU that I haven't been
able to solve, but I shall solve them in a later patch.
Pull Request: https://projects.blender.org/blender/blender/pulls/120135
The File Output node always has inputs of type Color regardless of the
type of the connected socket. This is because the socket type update
mechanism no longer works, so the default socket type persisted.
This is a regression due to 4a4916db45, where the RNA set function no
longer allows editing socket types for built-in nodes, and the File
Output node utilized the RNA set function. To fix this, we avoid the RNA
route and set the type directly.
This fix is related to #120175 and partially fixes it, though another
fix is still necessary.
Pull Request: https://projects.blender.org/blender/blender/pulls/120199
Avoid copying the positions array into the evaluated edit hints array
that's used to support editing with deformed positions when there is
a topology-changing procedural operation. In a simple test in sculpt
mode with 706k curve points, memory usage went from 78 to 70 MB.
This adds more duplication would be ideal, mainly because retrieving
the data with write access and making implicit sharing info for arbitrary
arrays aren't abstracted by implicit sharing utilities. It may be possible
to improve both of those aspects, either now or in the future.
Pull Request: https://projects.blender.org/blender/blender/pulls/120146
The Vector Blur node sometimes produces wrong output. That's because no
proper memory barriers existed for the SSBO written to by the velocity
dilation pass, so ensure a proper barrier exist.
The Vector Blur node has blocky artifacts. That's because one of the
buffers used in the velocity dilation pass was not zero initialized,
while it was written to using atomic max operations, so make sure it is
zero initialized.
This patch unifies the Defocus node between the CPU and GPU compositors.
Both nodes now use a variable sized bokeh kernel which is always odd
sized for a center pixel guarantee. Further the CPU implementation now
properly handles half pixel offsets when doing interpolation, and always
sets the threshold to zero similar to the GPU implementation.
The Defocus node maximum size option has no effect if Z Depth is not
enabled. That's because the code incorrectly uses the maximum CoC radius
even though it is not relevant if Z Depth is disabled. To fix this, use
the maximum size option directly in case of no Z Depth.
The GPU compositor Chroma Key node differs from the reference
implementation. This is caused by a typo, where the acceptance angle was
halved after computing tangent instead of before computing it.
There are still a few places that are more complicated where the replacement
to `IDP_New` isn't obvious, but this commit replaces most uses of the ugly
`IDPropertyTemplate` usage.
We already expose two other types that don't have a directly
corresponding socket type-- float2 and byte colors. This also solves
the inability to write to the nurbs_order attribute in geometry nodes.
Fixes#119843
Pull Request: https://projects.blender.org/blender/blender/pulls/119879
