Add an operator "Copy to Asset Library" for installing asset bundles
into already-existing asset libraries.
The operator is shown when:
- the "Current File" library is selected,
- the blend file name matches `*_bundle.blend`, and
- the file is not already located in an asset library.
The user can select a target asset library, then gets a "Save As"
dialogue box to select where in that library the file should be saved.
This allows for renaming, saving in a subdirectory, etc.
The Asset Catalogs from the asset bundle are merged into the target
asset library.
The operator will refuse to run when external files are referenced. This
is not done in its poll function, as it's quite an extensive operator
(it loops over all ID datablocks).
Reviewed by: Severin
Differential Revision: https://developer.blender.org/D13312
This enables calling code to deal with the blendfile handle themselves,
BKE_blendfile_link then just borrows, uses this handle and does not
release it.
Needed e.g. for python's libcontext system to use new
BKE_blendfile_link_append code.
Part of T91414: Unify link/append between WM operators and BPY context
manager API, and cleanup usages of `BKE_library_make_local`.
This enables calling code to deal with the blendfile handle themselves,
BKE_blendfile_link then just borrows, uses this handle and does not
release it.
Needed e.g. for python's libcontext system to use new
BKE_blendfile_link_append code.
Part of T91414: Unify link/append between WM operators and BPY context
manager API, and cleanup usages of `BKE_library_make_local`.
This was broken by rB0c3b215e7d5456878b155d13440864f49ad1f230.
The caching of loaded exr files needed some special treatment.
Differential Revision: https://developer.blender.org/D13313
This was probably broken by rB5c2330203e11e0d916960218b07d88d2193bf526.
For now just add the padding back in a spreadsheet specific way.
Differential Revision: https://developer.blender.org/D13315
There was a missing normals layer that was requested by the data transfer
modifier from the target object. The normal layer was correctly added to
the target object. However, it never reached the data transfer modifier
because the mesh was copied in `BKE_object_get_evaluated_mesh`
(in the call to `get_mesh_for_write`) and the copy does not include the normals
layer.
The solution is to not use `get_mesh_for_write` here which was only used
because `BKE_object_get_evaluated_mesh` returns a non-const `Mesh *`.
Mid term, it should actually return a `const Mesh *` to avoid the confusion.
Differential Revision: https://developer.blender.org/D13319
For some users, dropping assets into preview area may be more practical
due to space constraints or it may be just more intuitive.
Reviewed By: ISS
Differential Revision: https://developer.blender.org/D13311
For using the Outliner and/or the Asset Browser as scene independent
tools to organize a/v source material is necessary for the users to be
able to drag and drop data blocks into the VSE. This was also an
unfulfilled design target for the Outliner Gsoc project.
Datablocks won't be used directly. Path to file will be passed to strip
add operator instead.
Reviewed By: ISS
Differential Revision: https://developer.blender.org/D13304
Use the correct device function (hipDeviceGet) for multi GPU setups, instead
of hipGetDevice which just returns the default device.
Differential Revision: https://developer.blender.org/D13323
We already supported previews for lights, just didn't actually use them
when making a light object an asset. They were only used when making the
light data itself an asset.
Mistake in e7bea3fb6e.
We should only skip preview generation for objects that don't support
preview rendering, not completely forbid accessing preview data of such
IDs.
BVH2 triangle intersection was broken on the GPU since packed floats can't
be loaded directly into SSE. The better long term solution for performance
would be to build a BVH2 for GPU and Embree for CPU, similar to what we do
for OptiX.
Fix problem with duplicated initial character when initiating or
switching to new windows. This is done by updating our copies of state
and modes from the new window when it receives WM_IME_SETCONTEXT
message. This problem and fix are only for the Windows platform.
* Rename "Auto Tiles" to "Use Tiling", it's not really automatic and
confusing with the old auto tile size add-on.
* Rename "Adaptive" scrambling distance to "Automatic", to avoid confusion
with adaptive sampling.
Fix problem with duplicated initial character when initiating or
switching to new windows. This is done by updating our copies of state
and modes from the new window when it receives WM_IME_SETCONTEXT
message. This problem and fix are only for the Windows platform.
Add the following methods to the CurveEval class:
total_length() : returns the total length of the curve without needing to
allocate a new array
total_control_point_size() : returns the total number of control points without
needing to allocate a new array
* Rename "Auto Tiles" to "Use Tiling", it's not really automatic and
confusing with the old auto tile size add-on.
* Rename "Adaptive" scrambling distance to "Automatic", to avoid confusion
with adaptive sampling.
Use a map to speed up search for UI block names.
Time to redraw the node editor was decreased from
around 75-120ms to 40-70ms in a tree with many
Geometry Nodes.
Differential Revision: https://developer.blender.org/D13225
According to Blender selection rules, selections should be flushed
to containing elements. Added an EDMB_select_flush() after edit
mode booleans or intersects are done. Hopefully this doesn't break
any scripts that might have been depending on the old (broken) behavior.
Happens when device runs out of memory and Cycles is moving some
textures to the host memory.
The delayed memory free for OptiX BVH was moving data from one
device_memory to another, leaving the original device memory in
an invalid state. This was ruining the allocation map in the CUDA
device which is using pointer to the device_memory.
This change makes it so the memory pointer is stolen from BVH
into the delayed memory free list.
Additionally, forbid copying and moving instances of device_memory
and added sanity checks in the device implementation.
Differential Revision: https://developer.blender.org/D13316
This allows to reduce signature of several functions, and make it eaiser
to integrate more higher-level usages later on.
This should be a non-behavioral-change commit.
Part of T91414: Unify link/append between WM operators and BPY context
manager API, and cleanup usages of `BKE_library_make_local`.
This removes the last main usage of BLO's instantiation code.
Also required some limited refactoring of BKE_blendfile_link_append's
instantiation to make it more modular, and usable by both linking and
appending code paths.
NOTE: This should be a non-behavioral change commit.
Part of T91414: Unify link/append between WM operators and BPY context
manager API, and cleanup usages of `BKE_library_make_local`.
This will allow to expose all those advanced features of the WM
operators to other parts of the code, like the python library context
manager, copy/paste code, etc.
This is expected to be a strictly no-behavioral-change commit.
Part of T91414: Unify link/append between WM operators and BPY context
manager API, and cleanup usages of `BKE_library_make_local`.
Maniphest Tasks: T91414
Differential Revision: https://developer.blender.org/D13222
A good sphere preview material has a 1:1 UV ratio (so we see squares as
least distorted as possible), as well as good poles.
Square UV:
The original sphere expected a 2:1 panorama to be mapped to it. This
patch changes that (I scaled Y by 0.5) so that square textures look ok.
Poles:
The original sphere had a low initial resolution, so no ammount of
subdivision would fix the poles.
The sphere has a subdivision modifier with 0 resolution. Later (3.1?) I
want to try to change the resolution on-the-fly based on whether the material
has a displacement map.
Old sphere (1.9K vertices):
{F11845752, size=500px}
New sphere (2.0K vertices):
{F11845710, size=500px}
Differential Revision: https://developer.blender.org/D13309
`IMB_moviecache` is implemented as a singleton. When destructing the
singleton via `IMB_moviecache_destruct` it will not be created anymore
resulting inusage of unallocated memory and potentional memory
corruption.
When running blender this doesn't happen, but when creating images in
test cases the moviecache should be able to be recreated after it is
destroyed.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D13287
This patch separates the static-part from the dynamic-part when
generate brush masks. This makes the generation of brush masks 2-5 times
faster depending on the size of the brush.
More improvements can be done, this was just low hanging fruit.
This regression was introduced by D11832, but there was problems before
that as well. I seem to have missed it in review. See the differential
revision for a screenshot of the difference.
Differential Revision: https://developer.blender.org/D13216
It seems that update_from_old assumed there would be an old
tree view available in the old block. This works for the asset browser
because the tree is always drawn, but for the spreadsheet that depends
on having an active object, which isn't necessarily always true.
Differential Revision: https://developer.blender.org/D13301
Generating curve mask for 2d texture painting had some hard-coded
parameters that eventually weren't used in the algorithm (hardness and
rotation of the brush). This patch removes these parameters.
Curve mask generation is done during 2d texture painting. There are some
performance issues in this part of the code. Before addressing those we
move the code to CPP.
New drag&drop reordering code would call constraints reordering operator
with the generic context, and not the one from the panel's layout.
missing the "constraint" member which is mandatory for poll function to
properly deal with override vs. local constraints.
This commit fixes it by generating a temp bContextStore in the panel
re-ordering callback.
NOTE: this fix will have to be extended to modifiers (which happen to
work currently because they have an 'active' status), and gpencil
modifiers (which are also broken currently).
Differential Revision: https://developer.blender.org/D13291
The idea behind this change is the same as in
rB6ee2abde82ef121cd6e927995053ac33afdbb438.
A `MultiFunction::debug_parameter_name` method could be
added separately when necessary.
Previously, the function names were stored in `std::string` and were often
created dynamically (especially when the function just output a constant).
This resulted in a lot of overhead.
Now the function name is just a `const char *` that should be statically
allocated. This is good enough for the majority of cases. If a multi-function
needs a more dynamic name, it can override the `MultiFunction::debug_name`
method.
In my test file with >400,000 simple math nodes, the execution time improves from
3s to 1s.
This allows the use of better/more readable data structures,
and will also make some refactors to the search button easier.
The build completed on the buildbot for all platforms.
A recent security update to macOS 10.15.7 causes crashes when using Eevee and
various other 3D viewport features. It appears that glGenerateMipmap is
broken, causing a crash whenever its commands are flushed/submitted to the GPU.
Ideally this would be fixed in a driver update, however it's unlikely this will
happen. Earlier macOS versions have been receiving security updates for 2 years,
and that window has just passed for 10.15. Further, computers with these GPUs
can't upgrade to a newer macOS version.
As a workaround, disable mipmaps on these GPUs, by setting the mipmap max level
to 0 and not calling glGenerateMipmaps. Effects like depth of field also use
mipmaps, but fill in the mip levels by other means. In those cases we keep the
mipmap level.
Differential Revision: https://developer.blender.org/D13295
This fixes the the app crash happening when trying to render smoke as a dense
3D texture. The changes are related to matching up hipew with the actual HIP
headers.
Differential Revision: https://developer.blender.org/D13296
Same fix as rB0a3b4d4c64f1, but this time for greasepencil.
To repeat: dopesheet in greasepencil mode was ignoring the temporariy
visibility flag of collections. As a result, even though the dopesheet
was supposed to show animation data of visible greasepencils only was
still showing such data of greasepencils that were hidden by hiding
their collection.
Updates the Transfer Attributes and Capture Attributes nodes
to support attributes from instances.
Differential Revision: https://developer.blender.org/D13292
This patch removes a bunch of specific code for drawing the spreadsheet
data set region, which was an overly specific solution for a generic UI.
Nowadays, the UI tree view API used for asset browser catalogs is a much
better way to implement this behavior.
To make this possible, the tree view API is extended in a few ways.
Collapsibility can now be turned off, and whether an item should
be active is moved to a separate virtual function.
The only visual change is that the items are now drawn in a box,
just like the asset catalog.
Differential Revision: https://developer.blender.org/D13198
This isn't a problem in 3.0 or master, but I'm porting the spreadsheet
data set region to a tree view and ran into this. This line needs to
whether the function is empty before calling it.
Differential Revision: https://developer.blender.org/D13197
Now that `node_intern.hh` is a C++ header, we can use C++ types
there. This patch replaces the linked list of dragged links with a
vector. Also, the list of drag operator custom data, `nldrag`, doesn't
seem to need to be a list at all, so I just made it a unique pointer.
Differential Revision: https://developer.blender.org/D13252
Use the `GPU_SHADER_3D_POLYLINE_UNIFORM_COLOR` shader instead of `GPU_SHADER_2D_LINE_DASHED_UNIFORM_COLOR`.
This is just a partial solution as "protected" fcurves still use the dashed shader.
Differential Revision: https://developer.blender.org/D13290
CC-BY Blender Studio https://studio.blender.org
Update the splash artwork for the daily builds (`master`) to
celebrate the beginning of a new major series in Blender.
The badges of the Development Fund membership levels have been removed for
a simpler design, and the font matches the one used elsewhere in blender.org
After rB97533eede444217b, instances have their own attribute domain,
but the spreadsheet code worked under the assumption that the component
used the point domain. Old files have to re-select the instances data
source to make it properly active
Adds an attribute provider for instance attributes.
A new domain `ATTR_DOMAIN_INSTANCE` is implemented.
Instance attributes are not yet realized correctly.
Differential Revision: D13149
With very long ray distance, OptiX ends up traversing many BVH nodes due to
a feature that improves precision. However this causes very slow rendering.
We now avoid generating such long rays by rejecting the few samples that have
long ray distances and very low probability of being generated. This should not
meaningfully affect render results.
Thanks to Sergey and Patrick for the investigation.
After rBb9febb54a492, the evaluated mesh from a curve is now presented
to render engines as a separate mesh object, but some code still assumed
that a curve object itself could have an evaluated mesh. However, this is
still true for surface objects and metaballs, which don't
use geometry sets yet.
Differential Revision: https://developer.blender.org/D13272
Asset libraries should be directories on disk. By manually entering a
file path it was possible to have a single blend file as asset library,
but that was not a designed-for situation, and it doesn't play well
with the asset catalog system.
Original implementation was a quick prototype which should have never
landed as-is in master. It had very limiting constraints and did not
allow for any real further development.
This commit fixes the internal implementation to make more sensible,
maintainable and evolutive.
NOTE: This commit introduces another forward-incompatibility in the
Blender file format: Files saved after this commit won't open properly
in older versions of blender regarding local inserted constraints or
modifiers into overrides of linked data.
NOTE: Technical details: The 'anchor' item name/index is now stored in
`subitem_reference_` members, and the actual 'source' item name/index is
stored in `subitem_local_` members of the override property operation
data.
Previously, only the `subitem_local_` members were used, storing the
anchor item name/index, and assuming the 'source' item was always the
next in the list.
Milestone I of T82160.
Maniphest Tasks: T82160
Differential Revision: https://developer.blender.org/D13282
If a valid matching string is found, return that item, otherwise
fallback to the item matching the given index, if any.
This will be useful in RNA override code, and potentially other
areas where data in lists can be referenced by their names or indices.
The old code did not work when there were multiple ids with
the same name (which can happen when ids are linked in).
The solution is to use the session ids instead. Those are different
even when two ids have the same name.
Differential Revision: https://developer.blender.org/D11116
D13044 allowed the link color overlay to be used with custom sockets.
This no longer works due to a condition that checks if the socket is
standard or not, which was in place to avoid bad indexing of the
std_node_socket_colors array. Since that array is no longer used, this
condition needs to be removed.
Differential Revision: https://developer.blender.org/D13274
Reviewed By: Hans Goudey
Constraints, modifiers and NLA tracks can now report from RNA whether
they are defined as comming from the override's reference linked data,
or are local to the override.
This is a bit similar to rBb7260ca4c9f4b7618c9c214f1270e31d6ed9886b.
Sometimes a group node may not reference a node group
because it was linked and can't be found.
Separating geometry nodes into a new library will make it
easier to improve compile times with features like unity
builds and precompiled headers.
Differential Revision: https://developer.blender.org/D13261
Offsets for meta strip were invalid. No steps to reproduce the issue are
available, but it is quite possible that there are files with incorrect
state after issues with meta strips were fixed.
Ensure correct offsets for meta strips in versioning code.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D13257
The existing RNA CollectionRef class only offers a length() operation
which is sometimes used for checking if the collection is empty. This is
inefficient for certain collection types which do not have a native
length member; the entire list is iterated to find the count.
This patch creates an explicit empty() method to be used in such cases
for better semantics. Additionally, many collection types will behave
more efficiently when using the new method instead of checking length.
Making use of the new method will follow separately.
Differential Revision: https://developer.blender.org/D12314
Now with Geometry Nodes in Blender,
trees can become much bigger than before.
This changes the minimum zoom value in the node editor.
Differential Revision: https://developer.blender.org/D13254
Socket drawing can be heavy with many nodes.
This patch skips drawing them on scale < 0.2
when they are barely visible anyway.
Differential Revision: https://developer.blender.org/D13255
This feature has been desired for some time:
- https://rightclickselect.com/p/ui/Tqbbbc/allow-navigating-while-transforming (See comments);
- D1583;
- T37427;
In short, blocking navigation during transform limits the user to move the object only to visible areas within the screen and hinders the allocation of objects within closed meshes.
The node editor is also impaired because some nodes are far between them and the connectors are too small.
The only disadvantage of this patch (as I see it) is the conflict with the existing key map:
MIDDLEMOUSE:
- enable axis constrain in 3D view;
WHEELDOWNMOUSE, WHEELUPMOUSE, PAGEUPKEY, PAGEDOWNKEY:
- change the threshold of the proportional edit;
So the patch solution was to change these keymaps:
- MIDDLEMOUSE to Alt+MIDDLEMOUSE;
- WHEELDOWNMOUSE, WHEELUPMOUSE, PAGEUPKEY, PAGEDOWNKEY to Alt+(corresponding key);
When you use this new keymap for the first time in the proportional edit, it may seem strange due to the custom of using it (both in View2D and View3D).
But quickly the user gets used to it.
Alternatively we can add an option to the user preferences ([] Allow navigating while transforming). (I'm not much fan of this option).
The patch was done on branch2.8. But maybe it's a good idea to apply it to 2.79
Differential Revision: https://developer.blender.org/D2624
This patch implements part of what was stated in {T66484}, with respect to `Base Point`.
## Introduction
The snapping feature of the transform tools has a variety of applications:
- Organization of nodes.
- Positioning of frames in precise time units.
- Retopology with snap to face
- Creation of armatures with bone positioning through the snap to volume
- Precise positioning of 3D or 2D objects in the surrounding geometry (CAD modeling)
The goal of this document is to make it more powerful for precision modeling and still supporting the old use cases without extra complexity.
The main topic addressed here is the introduction of a **interactive mode for setting a snap source** (See terminology).
## Terminology
* **Snap Source**: 3d coordinate * we want to snap from. (Currently defined by the `Snap With` options: `Closest`, `Center`, `Median` and `Active`).
* **Snap Target**: 3d coordinate* we want to snap to. (Vertices, Edges, Faces, Grid...)
## Interactive Mode for Editing a Snap Source
Currently the fixed snap point can only be obtained through the `Snap With` options. So it's a little tricky for the user to define a snap source point having so much geometry on an object.
Because of this, the user needs to resort to impractical solutions to get a point in the geometry.
See example of an impractical use:
{F11714181, layout=left, width=960, alt="The user used the cursor (which can be snapped) to choose the snap origin point."}
The user used the cursor (which can be snapped) to choose the snap source point.
While it is possible to work around this current limitation, it is important to reduce the number of steps and allow the user to set a snap source point through an optional interactive mode during a transformation.
The proposed solution is to be able to move the current snap source point through a modal modifier activated with a key (eg. B).
The snap source point can thus "snap" to the elements in the scene (vertex, mid-edge, Lamp, …) during this mode.
{F9122814, layout=left, width=960, alt="Base Point Snap, example of transform operation via the shortcut (not the tool). After pressing g and the snap base change shortcut (e.g., shift + ctrl) the user set the base point. The base point is then visible until the end of the operation. The z axis constrains the final position."}
## Implementation Details
- The feature will only be available in 3D View.
- The feature will only be available for `Move`, `Rotate` and `Scale` transform modes.
- The snap source editing will be enabled with a single click on the modifier key (B).
- Having a snap point indicated, the new snap origin point will be confirmed with the same buttons that confirms the transformation (but the transformation will not be concluded).
- The snap source editing can be canceled with the same key that activated it (B).
- If the transformation is done with "release_confirm" (common for gizmos), the new feature cannot be enabled.
- During the transformation, when enabling the feature, if the snap option is turned off in the scene, the snap will be forced on throughout the rest of the transformation (unless interactive mode is canceled).
- During a transformation, if no snap target is set for an element in the scene (Vertex, Grid...), the snap targets to geometry Vertex, Edge, Face, Center of Edge and Perpendicular of Edge will be set automatically.
- Snap cannot be turned off during the snap source editing.
- Constraint or similar modification features will not be available during the snap source editing.
- Text input will not be available during the snap source editing.
- When adding multiple snap points (A) the new prone snap source point will be indicated with an "X" drawing.
{F11817267}
Maniphest Tasks: T66484
Differential Revision: https://developer.blender.org/D9415
Add the "Convert Old Pose Library" operator to the old pose library (in
the Armature properties editor). This makes it more discoverable; before
it only was available in the Action editor.
When filtering the asset browser, also include results that have partial
tag matches. So searching for "xite" will include results tagged with
"excited".
This brings the tag filtering in line with other search boxes in
Blender. Later we might want to provide users with more options for
prefix-only ("excite" would match "excited", but "xited" would not) or
only exact matches.
The Cycles accurate mode was removed, but the Eevee option for this has
a different meaning and should not have been removed. The Eevee accurate
makes cryptomatte accumulate for every sample, which Cycles has always
done regardless of any option.
MSL requires that constant address space literals be declared at program
scope. This patch moves the `blackbody_table_r/g/b` and `cie_colour_match`
constants into separate files so they can be declared at the appropriate scope.
Ref T92212
Differential Revision: https://developer.blender.org/D13241
This patch contains many small leftover fixes and additions that are
required for Metal-enablement:
- Address space fixes and a few other small compile fixes
- Addition of missing functionality to the Metal adapter headers
- Addition of various scattered `__KERNEL_METAL__` blocks (e.g. for
atomic support & maths functions)
Ref T92212
Differential Revision: https://developer.blender.org/D13263
Previously the check was done based on dimension of image and if any
of dimensions were larger than tile size tiling was used.
This change makes it so that if image does not exceed number of pixels
in the tile no tile will be used. Allows to render widescreen images
without tiling.
Differential Revision: https://developer.blender.org/D13206
The calculation based on preserving device occupancy was conflicting
with the fact that time limit needs to render less samples at the last
round of render work.
For example, rendering BMW27 for 30sec on i9-11900k was actually
rendering for almost a minute. Now the render time limit is respected
much more close.
Differential Revision: https://developer.blender.org/D13269
Build HIP kernels with NanoVDB, and patch NanoVDB to work with HIP.
This is a header only library so no rebuild is needed. The changes are being
submitted upstream to openvdb, so this patch should be temporary.
Thanks Thomas for help testing this.
For data-blocks from the current file, the image-buffer for dragging
wasn't set at all. This wasn't intentional, dragging things in the Asset
Browser should just always show the preview.
stack_assign_if was used in the middle of creating the shader value blocks.
Which caused stack variables to be inserted in the middle of the shader value data.
This resulted in the shader node data no being in sequential order. This was also
the case for the wave texture wave node.
Reviewed By: brecht
Maniphest Tasks: T93102
Differential Revision: https://developer.blender.org/D13262
This function can be used to "dissolve" meta strip anywhere in strip
hierarchy. This has same effect as `meta_separate` operator.
Reviewed By: ISS
Differential Revision: https://developer.blender.org/T91005
This patch changes the thumbnails' height used for image and movie
strips by choosing the proper size according to the VSE's text overlay
settings: i.e. thumbnails use the whole strip's height when no overlay
text is displayed; otherwise, some space is left for the overlay.
Reviewed By: ISS
Differential Revision: https://developer.blender.org/D13043
Some when adding multiple movies at once and only some of them have
audio track, this causes crash on NULL dereference. Issue was introduced
in bdbc7e12a0 to align sound and video properly.
Check if sound is present in movie file. If it's not, don't try to align
sound with video.
This patch adds a CMake option "WITH_CYCLES_DEBUG" which builds cycles with
a feature that allows debugging/selecting the direct-light sampling strategy.
The same option may later be used to add other debugging features that could
affect performance in release builds.
The three options are:
* Forward path tracing (e.g., via BSDF or phase function)
* Next-event estimation
* Multiple importance sampling combination of the previous two methods
Such a feature is useful for debugging light different sampling, evaluation,
and pdf methods (e.g., for light sources and BSDFs).
Differential Revision: https://developer.blender.org/D13152
Introduce a packed_float3 type for smaller storage that is exactly 3
floats, instead of 4. For computation float3 is still used since it can
use SIMD instructions.
Ref T92212
Differential Revision: https://developer.blender.org/D13243
Previously, node types had a callback that creates internal links. Pretty
much all nodes used the same callback though. The exceptions are the
reroute node (which probably shouldn't be mutable anyway) and some
input/output nodes that are not mutable.
Removing the callback helps with D13246, because it makes it easier
to reason about which internal links are created and when they change.
In the future, the internal links should be part of the node declaration.
This cleans up part of the code that still set the flag manually. Also, this
change helps with D13246 because it makes it easier to tag the node
tree as changed when the availability of a socket changed.
This patch includes:
View vector fix for ortho back face.
Point on segment logic correction.
Better handling of boundary cases.
See review page for detailed description.
Reviewed By: Sebastian Parborg (zeddb)
Differential Revision: https://developer.blender.org/D13143
This patch includes:
View vector fix for ortho back face.
Point on segment logic correction.
Better handling of boundary cases.
See review page for detailed description.
Reviewed By: Sebastian Parborg (zeddb)
Differential Revision: https://developer.blender.org/D13143
This change was introduced in 997b5fe45d, to not display pixelated
thumbnails. However when VSE timeline height is made smaller, this
limits strip height.
Change limit, so one strip can occupy full height of VSE timeline
This commit contains a few mostly-related changes to this node:
- Add a warning when retrieving the geometry from the modifier object.
- Only create the output geometry when it is necessary.
- Decompose transform matrices in a more friendly way.
- Use default return callbacks like other newer nodes.
Differential Revision: https://developer.blender.org/D13232
It's valid for a node group connected to the modifier not to
have a geometry input, but I didn't consider that case
with the last change I made here, f3bdabbe24.
Differential Revision: https://developer.blender.org/D13231
The spline code, especially Bezier splines, often make use of lazily
evaluation and caching. In order to do that, they use mutex locks.
When multi-threading, this can lead to problems. Further detail
can be found in rBfcc844f8fbd0d1.
To fix the deadlock, isolate the task before multi-threading
when holding a lock.
Differential Revision: https://developer.blender.org/D13229
Due to asynchronous process, the preview for a given image may be
generated several times.
This regenerates many thumbs unnecessarily.
The solution is to add the `FILE_ENTRY_PREVIEW_LOADING` flag for file
entries that are still in the thread queue.
So this flag is checked not to redraw the thumb when it is still being
created on a different thread.
Differential Revision: https://developer.blender.org/D11150
Fix a crash when a hair system's `ParticleSettings` ID datablock was
linked from another file but couldn't be found. This results in default
settings, with `type = PART_EMITTER`, where the particle data still has
a non-NULL `hair` pointer. Previously, copies of such a particle system
would NOT copy hair data for non-hair particle systems, hence the
pointer of the copy pointed to the original data, which got freed (at
least) twice upon closing the blend file.
This is now fixed by always copying the hair data, regardless of the
particle system type.
Reviewed by: mont29
Differential Revision: https://developer.blender.org/D13245
Add comment explaining `MEM_dupallocN` is NULL-safe, in that it returns
NULL when it receives a NULL pointer. This is currently true for both
implementations of the function (`MEM_lockfree_dupallocN` and
`MEM_guarded_dupallocN`), and will be expected of other implementations
as well.
No functional changes.
Add a new "experimental" debug option `show_asset_debug_info`, and use
that to determine the visibility of the active asset's catalog UUID and
simple name. Previously this was only determined by the "Developer
Extras" option, which meant it was visible in too many situations. It's
not really a "developer extra", and really just a debugging tool, so the
new option is more in line with its purpose.
Reviewed by: Severin
Differential Revision: https://developer.blender.org/D13242
This patch adapts the existing volumetric read/write lambda functions for Metal. Lambda expressions are not supported on MSL, so two new macros `VOLUME_READ_LAMBDA` and `VOLUME_WRITE_LAMBDA` have been defined with a default implementation which, on Metal, is overridden to use inline function objects.
This patch also removes the last remaining mention of the now-unused `ccl_addr_space`.
Ref T92212
Reviewed By: leesonw
Maniphest Tasks: T92212
Differential Revision: https://developer.blender.org/D13234
In 499dbb626a, the background color of drag tooltips were changed so
text becomes more readable. But multiple people were touching the same
code, so the disabled hint tooltips didn't get the same tweak. They
would benefit from them even more, since the red text is even harder to
read on the transparent background than the regular, white text.
Dispose of GHOST system paths when tearing down `BlendfileLoadingBaseTest`
and some other test cases. This prevents a memory leak.
A better solution would be to rework Blender's initialisation & teardown
structure, but that's outside the scope of this fix.
No functional changes to Blender.
While there is nothing technically that would cause issues when moving a
catalog into itself (it just changes the path of the catalog, and the
missing parent catalogs will be created), it seems broken to the user.
So disable this in the drag & drop code for asset catalogs.
Refreshing the assets requires `file_OT_asset_library_refresh` in the
asset browser, and `asset_OT_list_refresh` for the asset view. Both
are now done from `ASSET_OT_open_containing_blend_file`.
Python Operator templates made accessible from respective menus
(required to also use F3 search for quick access)
Also fixed Modal Draw Operator id_name (had duplicate name from other template)
Maniphest Tasks: T90866
Differential Revision: https://developer.blender.org/D13182
Goals of this refactor:
* Simplify creating virtual arrays.
* Simplify passing virtual arrays around.
* Simplify converting between typed and generic virtual arrays.
* Reduce memory allocations.
As a quick reminder, a virtual arrays is a data structure that behaves like an
array (i.e. it can be accessed using an index). However, it may not actually
be stored as array internally. The two most important implementations
of virtual arrays are those that correspond to an actual plain array and those
that have the same value for every index. However, many more
implementations exist for various reasons (interfacing with legacy attributes,
unified iterator over all points in multiple splines, ...).
With this refactor the core types (`VArray`, `GVArray`, `VMutableArray` and
`GVMutableArray`) can be used like "normal values". They typically live
on the stack. Before, they were usually inside a `std::unique_ptr`. This makes
passing them around much easier. Creation of new virtual arrays is also
much simpler now due to some constructors. Memory allocations are
reduced by making use of small object optimization inside the core types.
Previously, `VArray` was a class with virtual methods that had to be overridden
to change the behavior of a the virtual array. Now,`VArray` has a fixed size
and has no virtual methods. Instead it contains a `VArrayImpl` that is
similar to the old `VArray`. `VArrayImpl` should rarely ever be used directly,
unless a new virtual array implementation is added.
To support the small object optimization for many `VArrayImpl` classes,
a new `blender::Any` type is added. It is similar to `std::any` with two
additional features. It has an adjustable inline buffer size and alignment.
The inline buffer size of `std::any` can't be relied on and is usually too
small for our use case here. Furthermore, `blender::Any` can store
additional user-defined type information without increasing the
stack size.
Differential Revision: https://developer.blender.org/D12986
`ABCPointsWriter::is_supported` already checked for valid particle
system types (liquid, spray, foam, bubbles, ...).
`AbstractHierarchyIterator::make_writers_particle_systems` did not
create a writer for these though, so now bring these in line and also
create writers for these.
With transform tools, it is expected to see backgroud image when overlay
is transformed.
Alpha over caused performance to be not optimal when used with opaque
media. This should be addressed with D12914 at least partially. There
may be some corner cases not addressed.
Differential Revision: https://developer.blender.org/D12952
When scaling down image, users expect to see background, which doesn't
currently happen in VSE. This is because strips use cross blend mode by
default, because alpha over is much slower. Reason is, because any area
of image can be transparent, and therefore it can't have early out
implemented in a way that cross blend mode can.
Flag images rendered by codecs that don't support transparency as fully
opaque and implement a form of early out for alpha over blend mode.
When rendering image stack, 2-input effects are ignored on the
"way down". Alpha over needs rendered overlay image to decide whether it
will use only overlay or background too. Therefore overlay can be
rendered safely before it is used. Image flags can be checked and it can
be freed if needed. Freeing doesn't cause any performance degradation,
because image is always stored in cache.
This feature does not improve blend mode performance. In summary, it
only allowes for having alpha over blend mode on background images
without suffering from lower performance.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D12914
`v2d->tot` rect was set for backdrop drawing. Set range before drawing
scrollbars.
Reviewed By: Severin
Differential Revision: https://developer.blender.org/D13099
Some RNA properties and operators did not invalidate cache or did it
incorrectly.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D13101
Issue was caused by incorrect FFmpeg asynchronous decoding API. In most
cases, decoder returns 1 frame each time it is fed by 1 packet. Here
decoder wanted to return more frames, but our code always expected only
one.
Before sending new packets to decoder, check if there are frames to
receive. If there are, process them, otherwise continue decoding as
usual.
Reviewed By: zeddb, sergey
Differential Revision: https://developer.blender.org/D13079
When proxy size lower than 100% is used, clip strips are rendered with
incorrect image size.
This is because if proxies aren't enabled in movieclip, it automatically
falls back on rendering original media. Sequencer doesn't have knowledge
about this and since 9c99292a16 it assumes that image is proxy,
because it explicitly requested this size.
Check movieclip flag to see if proxies are enabled.
Reviewed By: sergey
Differential Revision: https://developer.blender.org/D13080
There were two issues:
- The third math node socket does not exist in old enough files.
- The comment incorrectly referred to the vector math node.
Differential Revision: https://developer.blender.org/D13219
When cut an stroke using the option Flat Caps, the falt was not done if the cut was done in the middle of the stroke.
Now the flat is applied to the segments created and also some cleanup of the code done.
Cause of this issue is that Custom Node Sockets info type was
initialized as SOCK_FLOAT when registering. Areas within the core that
would ignore custom socket types by checking its type would use the
socket as being a float type.
When custom node sockets have a property called default_value blender
tries to store it as an internal default value what failed in debug
builds.
This patch will set the socket type to SOCK_CUSTOM when registering a
custom socket type and allow, but skip storage of custom default values.
In this case the default values should already be stored as custom
properies.
Reviewed By: campbellbarton, JacquesLucke
Maniphest Tasks: T89260
Differential Revision: https://developer.blender.org/D13174
Allow the use of floating-point values for font point sizes, which
allows greater precision and flexibility for text output.
See D8960 for more information, details, and justification.
Differential Revision: https://developer.blender.org/D8960
Reviewed by Campbell Barton
Fix own mistake in rB7061d1e39fe
In my attempt to quickly address T92838, along with the original bug, I
made a nonsensical choice to use the limiter lock to guard the check
against the cache item itself. While harmless, it is not necessary and
semantically wrong / potentially confusing to future readers of the code.
Differential Revision: https://developer.blender.org/D13122
Suppressing the splash was only done when passing in an argument from
the command line.
Remove G.file_loaded, as it is misleading, only set once on startup,
replace with G.relbase_valid which is used everywhere else to check
if the file path should be used.
Support the ability to close relevant characters like '(', '[' and '{'.
It will also delete the pair character if they're empty.
Ref D13119
Reviewed By: campbellbarton
This reverts commit 6b4ca78108.
A simpler fix was used for 3.0, but rBd845ba481c6d2ef already contained
a more complete solution to the problem of inconsistent socket ids.
Previously, unique identifiers for sockets were created automatically,
sometimes using .001 and sometimes _001. Now they are created
manually with the second format, but some files were saved with .001
format. I think this was only an issue in the vector math node.
rBd845ba481c6d fixed this problem in 3.1, but in a more general way.
After I merge this patch to 3.1, I will revert it, since the versioning
added in that commit will make this redundant.
Differential Revision: https://developer.blender.org/D13209
The node does support curves, but only in index mode (see T88630)
So add a specific error message for the nearest mode, and let the
node support curves in the declaration.
Differential Revision: https://developer.blender.org/D13205
We need to increase GPU memory usage a bit. Unfortunately we can't get away
with writing either reflection or transmission passes because these BSDFs may
scatter in either direction but still must be in a fixed reflection or
transmission category to match up with the color passes.
Previously both `.` and `_` were used as separators when finding
a unique name for a socket. This removes the use of `.`, since `_`
was more common. It also does versioning for all of a file's node
trees to make sure that they all use the `_` convention.
Differential Revision: https://developer.blender.org/D13181
Partially reverts commit rB440a3475b8f5410e5c41bfbed5ce82771b41356f because
"optixDenoiserComputeIntensity" does not currently support input images that are not packed (the
"pixelStrideInBytes" field is not zero). As a result the intensity calculation would take into account
data from other passes in the image, some of which was scaled by the number of samples still and
therefore produce widely incorrect results that then caused artifacts in the denoised image.
Maniphest Tasks: T93029
The UI was always drawing all buttons in a layout, no matter if they
were scrolled out of view (as in, outside of the visible part of the
region) or not. This means it's doing quite some work that can be
avoided.
UI drawing generally isn't a big bottleneck in Blender, so I don't
expect huge speedups from this. But while playing back animation, we do
redraw a fair bit of the UI, so in cases where there are many buttons
out of view, it may bring a little FPS boost. E.g. say in complex node
trees (the node editor is redrawn on animation playback in case there
are animated values that need updated UI feedback). This also mitigates
the issue in T92922 significantly.
Differential Revision: https://developer.blender.org/T92922
Reviewed by: Brecht Van Lommel
UI_GetThemeColorBlendShade4fv incorrectly changing alpha by the amount
of the shading offset.
See D9944 for more details.
Differential Revision: https://developer.blender.org/D9944
Reviewed by Hans Goudey
UI_GetThemeColorBlendShade4fv incorrectly changing alpha by the amount
of the shading offset.
See D9944 for more details.
Differential Revision: https://developer.blender.org/D9944
Reviewed by Hans Goudey
Blender 3.0 will only support single-frame Actions in the pose library.
The goal of this patch is to lay the groundwork for making it possible
for the Asset Browser to reject/hide "animation snippet" Action assets.
Determining whether an Action has one or more frames (i.e. whether it
has a single pose or animation) requires inspecting the Action itself,
and thus loading the data-block itself. This would make it impossible to
quickly determine from the asset browser.
To solve this, the Action is inspected before saving, and a
`"is_single_frame"` boolean (well, 0/1 integer) IDProperty is added.
Reviewed by: Severin
Differential Revision: https://developer.blender.org/D13202
Channel protection flags were only used in global mode,
this doesn't make any sense, especially for rotation and scale.
Follow pose-bones, only using protection flags for
local & gimbal orientation.
Adds a boolean field output containing a selection of the
points of the star that are controlled by the outer radius
of the star.
Differential Revision: https://developer.blender.org/D13097
The root of the problem lies in bug in OIIO which we can work around
from our side (which does not affect pack memory usage).
Thanks Brecht for finding the root cause!
Differential Revision: https://developer.blender.org/D13186
When the file browser is in asset browser mode, it sets the callback
`filelist->prepare_filter_fn` to an asset browser specific function. This
function will segfault if there is no current asset library. Switching back
from asset browser to file browser would not reset that callback to
`NULL`, causing it to be called and crash Blender. This is now fixed.
* Rename the 'copy' functions to make it clear they belong to the same
'group' and are to be used together.
* Fix `flag` parameter of `BKE_copybuffer_paste` being a short instead
of an int.
* Improve documentation.
Adds a method to profiler that can be used to check if it is active.
This is used to determine if stop_profiling and start_profiling
should be called.
| patch | Juans Scene UI 256 samples | Juans Scene bg 256 samples | junkshop UI | junkshop bg |
| No patch | 6:16.59 | 4:05.37 | 2:08.48 | 1:59.7 |
| D13187 | 4:12.15 | 3:57.36 | 2:07.25 | 1:58.16 |
| D13185 | 4.11.18 |3:54.74 | 2:07.44 | 1:58.03 |
| D13190 | 4:12.39 | 3:55.42 | 2:07.62 | 1:58.68 |
UI - means rendered from within Blender
bg - means rendered from the command line using ##blender -b scene.blend -f 1##
Reviewed By: sergey, brecht
Maniphest Tasks: T92601
Differential Revision: https://developer.blender.org/D13190
Adds a method to profiler that can be used to check if it is active.
This is used to determine if stop_profiling and start_profiling
should be called.
| patch | Juans Scene UI 256 samples | Juans Scene bg 256 samples | junkshop UI | junkshop bg |
| No patch | 6:16.59 | 4:05.37 | 2:08.48 | 1:59.7 |
| D13187 | 4:12.15 | 3:57.36 | 2:07.25 | 1:58.16 |
| D13185 | 4.11.18 |3:54.74 | 2:07.44 | 1:58.03 |
| D13190 | 4:12.39 | 3:55.42 | 2:07.62 | 1:58.68 |
UI - means rendered from within Blender
bg - means rendered from the command line using ##blender -b scene.blend -f 1##
Reviewed By: sergey, brecht
Maniphest Tasks: T92601
Differential Revision: https://developer.blender.org/D13190
This adds a vulkan backend to GHOST. Still WIP, we should decide what
would be the minimum requirement to land in master.
Differential Revision: https://developer.blender.org/D13155
The mirror modifiers merge option caused unnecessary re-ordering
to the vertex array with original vertices merging into their copies.
While this wasn't an error, it meant creating a 1:1 mapping from input
vertices to their final output wasn't reliable (when looping over
vertices first to last) as is done in
BKE_editmesh_vert_coords_when_deformed.
As merging in either direction is supported, keep the source meshes
vertices in-order since it allows the vertex coordinates to be extracted.
The code assumed that any geometry input that wasn't the first input
was a second geometry input. Fix by separating the warning for the
first input and for the number of geometry inputs.
Adds a vector offset field to the "Curve Handle Position Node".
This vector is added to the incoming position (which is the
implicit handle position if not connected) which will set the
position of the handle. Default is (0,0,0)
Differential Revision: https://developer.blender.org/D13035
This patch adds list colors to the light theme for the spreadsheet,
which are needed for the data set region. Addresses T92492.
Differential Revision: https://developer.blender.org/D13090
Remove outdated CUDA comments for bindless textures and cleanup some HIP comments that still mentioned CUDA.
Differential Revision: https://developer.blender.org/D13189
When Constructing bezier splines from dna, the positions of the
left/right handles were set directly in the internal vectors, by
requesting a reference to them. The problem is that
BezierSpline::handle_positions_left() calls ensure_auto_handles()
before returning the reference. That function does some calculations on
uninitialized memory if the positions array is not yet filled.
Differential Revision: https://developer.blender.org/D13107
The issue was caused by splitting happening twice.
Fixed by checking for split flag which is assigned to the both states
during split.
The tricky part was to write catcher data at the moment of split: the
transparency and shadow catcher sample count is to be accumulated at
that point. Now it is happening in the `intersect_closest` kernel.
The downside is that render buffer is to be passed to the kernel, but
the benefit is that extra split bounce check is not needed now.
Had to move the passes write to shadow catcher header, since include
of `film/passes.h` causes all the fun of requirement to have BSDF
data structures available.
Differential Revision: https://developer.blender.org/D13177
rB60fee69682ac39 only partially fixed the issue, `BlendFileReadReport
bf_reports` was now properly stored in `BPy_Library` `self` for the
lifetime of the context, but its `reports` member was still referencing
local variable to `bpy_lib_enter` function.
This patch exposes the sampling offset option to Blender. It is located in the "Sampling > Advanced" panel.
For example, this can be useful to parallelize rendering and distribute different chunks of samples for each computer to render.
---
I also had to add this option to `RenderWork` and `RenderScheduler` classes so that the sample count in the status string can be calculated correctly.
Reviewed By: leesonw
Differential Revision: https://developer.blender.org/D13086
Apply a local-workaround instead of adding support for this use-case
since pre-selection isn't the intended purpose of gizmos.
This also resolves a glitch where poly-build and loop cut would
briefly show loop-cut or poly-build pre-selection after transforming.
See gizmo_preselect_poll_for_draw note for more details.
#### Motivation
The View pie menu is a convenient way to access operators such as `Frame Selected` and `Frame All` which are usually mapped to `PERIOD` or `HOME` keys on the right side of most keyboard, making it hard hard to reach with the left hand.
The motivation for this patch comes from working with a 75% keyboard (no numpad). Most laptops face a similar problem.
#### Implementation
The View pie menu has been added to the following editors and sub-modes where applicable:
* Node Editor
* Video Sequencer
* Dopesheet
* Graph
* NLA
* Image
* Clip
* Outliner
More options could definitely be added to this menu for convenience, as long as it maintains the common options in the same place (Frame Selected on the left, Frame All on the right).
For positioning I went with the following layout:
{F11791186, size=full}
I've added `Zoom 1:1`to the Image Editor and the VSE Preview since there is no way to reset the zoom on keyboards without numpad (unless Emulate Numpad is turned on).
The Outliner uses `Show Active` and `Show Hierarchy` which are the closest ones to the equivalent in other editors. Should `Show Active` be renamed to `Frame Selected`?
The shortcut assigned is the same as the 3D Viewport (`ACCENT_GRAVE`).
#### Screenshots
Node Editor
{F11778387, size=full}
Dopesheet
{F11778400, size=full}
Graph
{F11778403, size=full}
Image Editor (Paint and View)
{F11791113, size=full}
Image Editor (Mask)
{F11791114, size=full}
UV Editor
{F11791119, size=full}
Clip Editor (Tracking)
{F11791137, size=full}
Clip Editor (Mask)
{F11791140, size=full}
Clip Editor (Graph)
{F11791151, size=full}
View operators are not yet implemented in Clip Editor Dopesheet mode (left a note about this in the menu poll).
Reviewed By: #user_interface, campbellbarton
Differential Revision: https://developer.blender.org/D13169
rBc473b2ce8bdbf8fa42 improved the situation somewhat, but
attribute search still crashes during animation playback, because
the UI search data references stale memory. The proper solution
is to allow the search to own data rather than just referencing it,
but I would prefer not to do that for 3.0. In the meantime, just
disable attribute search when animation is playing.
Differential Revision: https://developer.blender.org/D13179
This is due to a driver bug, so disable it for now until it gets resolved
in a future driver release.
Ref T92972
Differential Revision: https://developer.blender.org/D13167
It's unclear why this fails. Maybe the size of half4 is not the expected
8 bytes and adjacent pixels are overwritten. Or there is some bug in the
HIP compiler writing a struct into global memory, which we probably don't
do elsewhere in the kernel.
Thanks to Thomas, William and Jeroen for helping investigate this.
This polishes up the build support for WITH_VULKAN
the find_package calls were moved to the appropriate
place in the platform cmake files.
For windows the required binaries are copied to
the binary folder.
Windows support relies on D11678 landing and
libraries being added to SVN and will automatically
disable WITH_VULKAN when they are not found.
The ram being uninitialised caused issues in
VKTexture's destructor which tests for
VK_NULL_HANDLE. The nvidia driver did not
enjoy being fed a bogus pointer and crashed.
Retrieve device platform information for `GPUPlatformGlobal`.
As vulkan needs a physical device it was moved from the GPUBackend
constructor to an init method that is called when an initial context
is created.
Still need to find out how to extract the driver type.
For the vulkan backend we control the compiler and therefore the log
parsing is much simpler then the OpenGL parsers.
This patch formats the log to a better readable/debuggable format so
developers can quickly detect what needs to be done.
Migration of the shaders can take a long time. During this time it
you won't be able to start most GPU tests or Blender. This switch
will enable the compilation of the shader to other developments
can still happen.
`WITH_VULKAN_SHADER_COMPILATION` Default is OFF. Enable when migrating
shaders.
Shaders are expected to fail as they need to be converted to Vulkan. Dy default the compilation is turned off. Enable VULKAN_SHADER_COMPILATION to activate shader module compilation.
The shader_compiler is an abstration layer for in case we need to switch
between other shader compilers in the future.
The shader compiler uses a CPP Api as it will only be accessed from GPU
module.
```
const char *source = "#version 450\nvoid main() {}";
shader_compiler::Compiler *compiler = shader_compiler::Compiler::create_default();
shader_compiler::Job job;
job.source = source;
job.name = __func__;
job.source_type = shader_compiler::SourceType::GlslComputeShader;
job.compilation_target = shader_compiler::TargetType::SpirV;
shader_compiler::Result *result = compiler->compile(job);
EXPECT_EQ(result->type, shader_compiler::TargetType::SpirV);
EXPECT_EQ(result->status_code, shader_compiler::StatusCode::Ok);
EXPECT_GT(result->bin.size(), 0);
EXPECT_EQ(result->error_log, "");
delete result;
delete compiler;
```
The CMakeFiles need some attention as it currently works when linking to
the shaderlib_combined. On linux it picks up the shaderlib from the
vulkan SDK, eventually it needs to pick up from our prebuild libraries.
On linux both the display and window should be available to create a
window surface. Only the display was checked making it fail later on
when creating a present queue.
When using a dense mesh and transforming a small number of verts the mesh received a copy-on-write signal. This will free all GPU batches. No reuse is currently possible.
This patch adds a test case that can be used as a baseline to optimize the cache construction in the draw module.
Differential Revision: https://developer.blender.org/D11255
This removes designated initializer because not supported by MSVC for our
current C++ version.
Fixes some other compile errors and add usage of the `ContextVK` context in
the Win32 backend.
The creator arg is meant to support more than vulkan or opengl.
This currently fails because there is no GPU backend separation.
Note that we use G.debug with a debug flag to know what context type to
create. This will need to be cleaned later.
option(WITH_VULKAN_SHADER_COMPILATION"Temporary flag to enable vulkan shader compilation needed to continue development during the migration of GLSL to Vulkan."OFF)
# OpenGL
option(WITH_OPENGL"When off limits visibility of the opengl headers to just bf_gpu and gawain (temporary option for development purposes)"ON)
option(WITH_OPENGL"When off limits visibility of the opengl headers to just bf_gpu (temporary option for development purposes)"ON)
option(WITH_GLEW_ES"Switches to experimental copy of GLEW that has support for OpenGL ES. (temporary option for development purposes)"OFF)
option(WITH_GL_EGL"Use the EGL OpenGL system library instead of the platform specific OpenGL system library (CGL, glX, or WGL)"OFF)
option(WITH_GL_PROFILE_ES20"Support using OpenGL ES 2.0. (through either EGL or the AGL/WGL/XGL 'es20' profile)"OFF)
@@ -521,6 +530,7 @@ mark_as_advanced(
WITH_GLEW_ES
WITH_GL_EGL
WITH_GL_PROFILE_ES20
WITH_VULKAN_SHADER_COMPILATION
)
if(WIN32)
@@ -1065,7 +1075,7 @@ if(MSVC)
add_definitions(-D__LITTLE_ENDIAN__)
# OSX-Note: as we do cross-compiling with specific set architecture,
# endianess-detection and auto-setting is counterproductive
# endianness-detection and auto-setting is counterproductive
# so we just set endianness according CMAKE_OSX_ARCHITECTURES
Easy to integrate Vulkan memory allocation library.
**Documentation:** See [Vulkan Memory Allocator](https://gpuopen-librariesandsdks.github.io/VulkanMemoryAllocator/html/) (generated from Doxygen-style comments in [src/vk_mem_alloc.h](src/vk_mem_alloc.h))
**License:** MIT. See [LICENSE.txt](LICENSE.txt)
**Changelog:** See [CHANGELOG.md](CHANGELOG.md)
**Product page:** [Vulkan Memory Allocator on GPUOpen](https://gpuopen.com/gaming-product/vulkan-memory-allocator/)
Memory allocation and resource (buffer and image) creation in Vulkan is difficult (comparing to older graphics API-s, like D3D11 or OpenGL) for several reasons:
- It requires a lot of boilerplate code, just like everything else in Vulkan, because it is a low-level and high-performance API.
- There is additional level of indirection: `VkDeviceMemory` is allocated separately from creating `VkBuffer`/`VkImage` and they must be bound together.
- Driver must be queried for supported memory heaps and memory types. Different IHVs provide different types of it.
- It is recommended practice to allocate bigger chunks of memory and assign parts of them to particular resources.
# Features
This library can help game developers to manage memory allocations and resource creation by offering some higher-level functions:
1. Functions that help to choose correct and optimal memory type based on intended usage of the memory.
- Required or preferred traits of the memory are expressed using higher-level description comparing to Vulkan flags.
2. Functions that allocate memory blocks, reserve and return parts of them (`VkDeviceMemory` + offset + size) to the user.
- Library keeps track of allocated memory blocks, used and unused ranges inside them, finds best matching unused ranges for new allocations, respects all the rules of alignment and buffer/image granularity.
3. Functions that can create an image/buffer, allocate memory for it and bind them together - all in one call.
Additional features:
- Well-documented - description of all functions and structures provided, along with chapters that contain general description and example code.
- Thread-safety: Library is designed to be used in multithreaded code. Access to a single device memory block referred by different buffers and textures (binding, mapping) is synchronized internally.
- Configuration: Fill optional members of CreateInfo structure to provide custom CPU memory allocator, pointers to Vulkan functions and other parameters.
- Customization: Predefine appropriate macros to provide your own implementation of all external facilities used by the library, from assert, mutex, and atomic, to vector and linked list.
- Support for memory mapping, reference-counted internally. Support for persistently mapped memory: Just allocate with appropriate flag and you get access to mapped pointer.
- Support for non-coherent memory. Functions that flush/invalidate memory. `nonCoherentAtomSize` is respected automatically.
- Support for resource aliasing (overlap).
- Support for sparse binding and sparse residency: Convenience functions that allocate or free multiple memory pages at once.
- Custom memory pools: Create a pool with desired parameters (e.g. fixed or limited maximum size) and allocate memory out of it.
- Linear allocator: Create a pool with linear algorithm and use it for much faster allocations and deallocations in free-at-once, stack, double stack, or ring buffer fashion.
- Support for Vulkan 1.0, 1.1, 1.2.
- Support for extensions (and equivalent functionality included in new Vulkan versions):
- VK_EXT_memory_budget: Used internally if available to query for current usage and budget. If not available, it falls back to an estimation based on memory heap sizes.
- VK_KHR_dedicated_allocation: Just enable it and it will be used automatically by the library.
- VK_AMD_device_coherent_memory
- VK_KHR_buffer_device_address
- Defragmentation of GPU and CPU memory: Let the library move data around to free some memory blocks and make your allocations better compacted.
- Lost allocations: Allocate memory with appropriate flags and let the library remove allocations that are not used for many frames to make room for new ones.
- Statistics: Obtain detailed statistics about the amount of memory used, unused, number of allocated blocks, number of allocations etc. - globally, per memory heap, and per memory type.
- Debug annotations: Associate string with name or opaque pointer to your own data with every allocation.
- JSON dump: Obtain a string in JSON format with detailed map of internal state, including list of allocations and gaps between them.
- Convert this JSON dump into a picture to visualize your memory. See [tools/VmaDumpVis](tools/VmaDumpVis/README.md).
- Debugging incorrect memory usage: Enable initialization of all allocated memory with a bit pattern to detect usage of uninitialized or freed memory. Enable validation of a magic number before and after every allocation to detect out-of-bounds memory corruption.
- Record and replay sequence of calls to library functions to a file to check correctness, measure performance, and gather statistics.
# Prequisites
- Self-contained C++ library in single header file. No external dependencies other than standard C and C++ library and of course Vulkan. STL containers are not used by default.
- Public interface in C, in same convention as Vulkan API. Implementation in C++.
- Error handling implemented by returning `VkResult` error codes - same way as in Vulkan.
- Interface documented using Doxygen-style comments.
- Platform-independent, but developed and tested on Windows using Visual Studio. Continuous integration setup for Windows and Linux. Used also on Android, MacOS, and other platforms.
# Example
Basic usage of this library is very simple. Advanced features are optional. After you created global `VmaAllocator` object, a complete code needed to create a buffer may look like this:
3. An unused region of the memory block is bound to this buffer.
`VmaAllocation` is an object that represents memory assigned to this buffer. It can be queried for parameters like Vulkan memory handle and offset.
# Binaries
The release comes with precompiled binary executables for "VulkanSample" application which contains test suite and "VmaReplay" tool. They are compiled using Visual Studio 2019, so they require appropriate libraries to work, including "MSVCP140.dll", "VCRUNTIME140.dll", "VCRUNTIME140_1.dll". If their launch fails with error message telling about those files missing, please download and install [Microsoft Visual C++ Redistributable for Visual Studio 2015, 2017 and 2019](https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads), "x64" version.
# Read more
See **[Documentation](https://gpuopen-librariesandsdks.github.io/VulkanMemoryAllocator/html/)**.
# Software using this library
- **[Detroit: Become Human](https://gpuopen.com/learn/porting-detroit-3/)**
- **[Vulkan Samples](https://github.com/LunarG/VulkanSamples)** - official Khronos Vulkan samples. License: Apache-style.
- **[Anvil](https://github.com/GPUOpen-LibrariesAndSDKs/Anvil)** - cross-platform framework for Vulkan. License: MIT.
- **[Filament](https://github.com/google/filament)** - physically based rendering engine for Android, Windows, Linux and macOS, from Google. Apache License 2.0.
- **[Atypical Games - proprietary game engine](https://developer.samsung.com/galaxy-gamedev/gamedev-blog/infinitejet.html)**
- **[Flax Engine](https://flaxengine.com/)**
- **[Lightweight Java Game Library (LWJGL)](https://www.lwjgl.org/)** - includes binding of the library for Java. License: BSD.
- **[PowerVR SDK](https://github.com/powervr-graphics/Native_SDK)** - C++ cross-platform 3D graphics SDK, from Imagination. License: MIT.
- **[Skia](https://github.com/google/skia)** - complete 2D graphic library for drawing Text, Geometries, and Images, from Google.
- **[vkDOOM3](https://github.com/DustinHLand/vkDOOM3)** - Vulkan port of GPL DOOM 3 BFG Edition. License: GNU GPL.
- **[vkQuake2](https://github.com/kondrak/vkQuake2)** - vanilla Quake 2 with Vulkan support. License: GNU GPL.
- **[Vulkan Best Practice for Mobile Developers](https://github.com/ARM-software/vulkan_best_practice_for_mobile_developers)** from ARM. License: MIT.
- **[RPCS3](https://github.com/RPCS3/rpcs3)** - PlayStation 3 emulator/debugger. License: GNU GPLv2.
[Many other projects on GitHub](https://github.com/search?q=AMD_VULKAN_MEMORY_ALLOCATOR_H&type=Code) and some game development studios that use Vulkan in their games.
- **[Awesome Vulkan](https://github.com/vinjn/awesome-vulkan)** - a curated list of awesome Vulkan libraries, debuggers and resources.
- **[VulkanMemoryAllocator-Hpp](https://github.com/malte-v/VulkanMemoryAllocator-Hpp)** - C++ binding for this library. License: CC0-1.0.
- **[PyVMA](https://github.com/realitix/pyvma)** - Python wrapper for this library. Author: Jean-Sébastien B. (@realitix). License: Apache 2.0.
- **[vk-mem](https://github.com/gwihlidal/vk-mem-rs)** - Rust binding for this library. Author: Graham Wihlidal. License: Apache 2.0 or MIT.
- **[Haskell bindings](https://hackage.haskell.org/package/VulkanMemoryAllocator)**, **[github](https://github.com/expipiplus1/vulkan/tree/master/VulkanMemoryAllocator)** - Haskell bindings for this library. Author: Joe Hermaszewski (@expipiplus1). License BSD-3-Clause.
- **[vma_sample_sdl](https://github.com/rextimmy/vma_sample_sdl)** - SDL port of the sample app of this library (with the goal of running it on multiple platforms, including MacOS). Author: @rextimmy. License: MIT.
- **[vulkan-malloc](https://github.com/dylanede/vulkan-malloc)** - Vulkan memory allocation library for Rust. Based on version 1 of this library. Author: Dylan Ede (@dylanede). License: MIT / Apache 2.0.
('ACCURATE',"Accurate","Prefilter noisy guiding passes before denoising color. Improves quality when guiding passes are noisy using extra processing time",3),
)
enum_direct_light_sampling_type=(
('MULTIPLE_IMPORTANCE_SAMPLING',"Multiple Importance Sampling","Multiple importance sampling is used to combine direct light contributions from next-event estimation and forward path tracing",0),
('FORWARD_PATH_TRACING',"Forward Path Tracing","Direct light contributions are only sampled using forward path tracing",1),
('NEXT_EVENT_ESTIMATION',"Next-Event Estimation","Direct light contributions are only sampled using next-event estimation",2),
)
defupdate_render_passes(self,context):
scene=context.scene
view_layer=context.view_layer
@@ -325,6 +331,13 @@ class CyclesRenderSettings(bpy.types.PropertyGroup):
default=1024,
)
sample_offset:IntProperty(
name="Sample Offset",
description="Number of samples to skip when starting render",
min=0,max=(1<<24),
default=0,
)
time_limit:FloatProperty(
name="Time Limit",
description="Limit the render time (excluding synchronization time)."
@@ -346,7 +359,7 @@ class CyclesRenderSettings(bpy.types.PropertyGroup):
name="Scrambling Distance",
default=1.0,
min=0.0,max=1.0,
description="Lower values give faster rendering with GPU rendering and less noise with all devices at the cost of possible artifacts if set too low. Only works when not using adaptive sampling",
description="Reduce randomization between pixels to improve GPU rendering performance, at the cost of possible rendering artifacts if set too low. Only works when not using adaptive sampling",
)
preview_scrambling_distance:BoolProperty(
name="Scrambling Distance viewport",
@@ -354,10 +367,10 @@ class CyclesRenderSettings(bpy.types.PropertyGroup):
description="Uses the Scrambling Distance value for the viewport. Faster but may flicker",
)
adaptive_scrambling_distance:BoolProperty(
name="Adaptive Scrambling Distance",
auto_scrambling_distance:BoolProperty(
name="Automatic Scrambling Distance",
default=False,
description="Uses a formula to adapt the scrambling distance strength based on the sample count",
description="Automatically reduce the randomization between pixels to improve GPU rendering performance, at the cost of possible rendering artifacts. Only works when not using adaptive sampling",
)
use_layer_samples:EnumProperty(
@@ -415,6 +428,13 @@ class CyclesRenderSettings(bpy.types.PropertyGroup):
default=0,
)
direct_light_sampling_type:EnumProperty(
name="Direct Light Sampling Type",
description="The type of strategy used for sampling direct light contributions",
items=enum_direct_light_sampling_type,
default='MULTIPLE_IMPORTANCE_SAMPLING',
)
min_light_bounces:IntProperty(
name="Min Light Bounces",
description="Minimum number of light bounces. Setting this higher reduces noise in the first bounces, "
@@ -770,8 +790,8 @@ class CyclesRenderSettings(bpy.types.PropertyGroup):
)
use_auto_tile:BoolProperty(
name="Auto Tiles",
description="Automatically render high resolution images in tiles to reduce memory usage, using the specified tile size. Tiles are cached to disk while rendering to save memory",
name="Using Tiling",
description="Render high resolution images in tiles to reduce memory usage, using the specified tile size. Tiles are cached to disk while rendering to save memory",
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
