Replaces current ID Mask node Anti-Aliasing operation by SMAA
operations with default settings as proposed by Jeroen Bakker.
SMAA produces smoother edges.
Reviewed By: manzanilla
Differential Revision: https://developer.blender.org/D11881
Adds full frame implementation to this node operations.
No functional changes.
2.5x faster than tiled fallback on average.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11690
Not causing issues in current master because all buffer areas are at
(0, 0) position and `Extend` is not used. But areas may be at any
position in future developments and it will crash.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11784
Link sockets are always connected to inserted translate or scale
operation `Color` sockets even when they have different data type.
This causes crashes on full frame mode when operations read inputs
with non expected datatypes.
Because data type conversions need to be executed before, convert
resolutions must ensure same datatypes are linked.
Work scheduler needed initialization and execution models are
not created during constant folding. This moves work execution
method to execution system.
Graphs are usually large, needing a lot of horizontal scrolling and
they can include more information for debugging.
This patch makes graph more compact horizontally by splitting
labels in lines and removing namespaces.
Furthermore it adds following information:
- Operation ID.
- SetValueOperation float value.
- Optionally, operation node name.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11720
Currently there is no clear way to know if an operation is constant
(i.e. when all rendered pixels have same values). Operations may
need to get constant input values before rendering to determine
their resolution or areas of interest. This is the case of scale, rotate
and translate operations. Only "set operations" are known as
constant but many more are constant when all their inputs are so.
Such cases can be optimized by only rendering one pixel.
Current solution for tiled implementation is to get first pixel
from input. This works for root execution groups, others
need previous groups to be rendered.
On full frame implementation this is not possible, because buffers
are created on rendering to reduce peak memory and there is
no per pixel calls.
This patch evaluates all operations that are constant into primitive
operations (Value/Vector/Color) before determining resolutions.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11490
Adds full frame implementation to this node operations.
No functional changes.
1.3x faster than tiled fallback.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11764
Adds full frame implementation to this node operation.
No functional changes.
1.4x faster than tiled fallback.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11765
Adds full frame implementation to this node operation.
No functional changes.
1.7x faster than tiled fallback.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11766
Adds full frame implementation to this node operation.
No functional changes.
1.5x faster than tiled fallback.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11767
Adds full frame implementation to this node operations.
No functional changes.
1.2x faster than tiled fallback.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11638
Adds full frame implementation to Image node operations.
Mostly refactored into buffer utility methods for reuse in other
operations.
No functional changes.
1.8x faster than tiled fallback.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11559
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
On selecting a multi-layer image with a combined pass, a "Combined"
socket is created and default combined pass socket "Image" is
disabled by setting `SOCK_UNAVAIL` flag. When converting into
operations, `ImageNode` converts alpha socket on finding any socket with
a combined pass without checking the flag.
Since commit rB93e2491ee724 an assertion fails when mapping sockets
twice because now map `add_new` is used.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11566
Adds full frame implementation to Value node operation.
No functional changes.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11594
After looking into task isolation issues with Sergey, we couldn't find the
reason behind the deadlocks that we are getting in T87938 and a Sprite Fright
file involving motion blur renders.
There is no apparent place where we adding or waiting on tasks in a task group
from different isolation regions, which is what is known to cause problems. Yet
it still hangs. Either we do not understand some limitation of TBB isolation,
or there is a bug in TBB, but we could not figure it out.
Instead the idea is to use isolation only where we know we need it: when
holding a mutex lock and then doing some multithreaded operation within that
locked region. Three places where we do this now:
* Generated images
* Cached BVH tree building
* OpenVDB lazy grid loading
Compared to the more automatic approach previously used, there is the downside
that it is easy to miss places where we need isolation. Yet doing it more
automatically is also causing unexpected issue and bugs that we found no
solution for, so this seems better.
Patch implemented by Sergey and me.
Differential Revision: https://developer.blender.org/D11603
Changes introduced in commit rBe9f2f17e8518
can create different render results when there is
a Math or Mix operation after TextureOperation
on tiled execution model.
This is due to WriteBufferOperation forcing a single pixel
resolution when these operations use a preferred
resolution of 0 to check if their inputs have resolution.
Fixing this behaviour creates different renders too.
This patch keeps previous tiled implementation and
adds the new implementation only for full frame execution.
Reviewed By: Jeroen Bakker (jbakker)
Differential Revision: https://developer.blender.org/D11546
In order to reduce stack size this patch converts full frame
recursive methods into iterative.
- No functional changes.
- No performance changes.
- Memory peak may slightly vary depending on the tree because
now breadth-first traversal is used instead of depth-first.
Tests in D11113 have same results except for test1 memory peak:
360MBs instead of 329.50MBs.
Reviewed By: Jeroen Bakker (jbakker)
Differential Revision: https://developer.blender.org/D11515
Under some circumstances using task isolation can cause deadlocks.
Previously, our task pool implementation would run all tasks in an
isolated region. Now using task isolation is optional and can be
turned on/off for individual task pools.
Task pools that spawn new tasks recursively should never enable
task isolation. There is a new check that finds these cases at runtime.
Right now this check is disabled, so that this commit is a pure refactor.
It will be enabled in an upcoming commit.
This fixes T88598.
Differential Revision: https://developer.blender.org/D11415
This patch adds the base code needed to make the full-frame system work for both current tiled/per-pixel implementation of operations and full-frame.
Two execution models:
- Tiled: Current implementation. Renders execution groups in tiles from outputs to input. Not all operations are buffered. Runs the tiled/per-pixel implementation.
- FullFrame: All operations are buffered. Fully renders operations from inputs to outputs. Runs full-frame implementation of operations if available otherwise the current tiled/per-pixel. Creates output buffers on first read and free them as soon as all its readers have finished, reducing peak memory usage of complex/long trees. Operations are multi-threaded but do not run in parallel as Tiled (will be done in another patch).
This should allow us to convert operations to full-frame in small steps with the system already working and solve the problem of high memory usage.
FullFrame breaking changes respect Tiled system, mainly:
- Translate, Rotate, Scale, and Transform take effect immediately instead of next buffered operation.
- Any sampling is always done over inputs instead of last buffered operation.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11113
Cause is that initializing the cryptomatte session would reset the
current frame of an image sequence. The solution is to always use the
scene current frame so it resets to the correct frame.
This was a todo that wasn't solved after it landed in master.
Needs to be backported to 2.93.
When compositor node tree has a texture node, TextureOperation vector inputs has always {0, 0} resolution instead of having same resolution as TextureOperation which is the expected behaviour for resolutions propagation.
Current TextureOperation determineResolution implementation doesn't determine inputs resolution, breaking propagation of preferred resolution and that's the reason why they are always 0. Setting scene resolution always would mean it is its own resolution and could make sense, but setting it only when preferred resolution is 0, breaks preferred resolution logic affecting other operations as explained in D10972. In any case scene resolution is already the default preferred resolution on viewer and compositor nodes.
Reviewed By: jbakker
Differential Revision: https://developer.blender.org/D11381
Deduplicates code by introducing a PlaneDirtortBaseOperation for common logic.
Reviewed By: #compositing, jbakker
Differential Revision: https://developer.blender.org/D11273
