While this preprocessing does take some time upfront,
it avoids longer lookup later on, especially as nodes get
more sockets.
It's probably possible to make this more efficient in some cases
but this is good enough for now.
Scanline processor did its own heurestic what didn't scale well when
having a multiple cores. In stead of using our own code this patch will
leave it to TBB to determine how to split the scanlines over the
available threads.
Performance of the IMB_transform before this change was 0.002123s, with
this change 0.001601s. This change increases performance in other areas
as well including color management conversions.
Reviewed By: zeddb
Differential Revision: https://developer.blender.org/D11578
When using large sequences including audio the drawing of the audio on
top of the strip takes a lot of time. This effects the playback
performance heavily.
During the animation playback performance there was a solution for this
by only drawing the playhead overlay. This was reverted for the sequence
editor as it didn't update the color strips when they were animated.
This patch checks if there are animated color strips if so the full
screen is redrawn, otherwise only the playhead is redrawn.
Reviewed By: ISS
Differential Revision: https://developer.blender.org/D11580
One drawback to trying to predict the number of threads that will be
used in the `task_graph` is that we are only sure of the number when the
threads are running.
Using `BLI_task_parallel_range` allows the driver to
choose the best thread distribution through `parallel_reduce`.
The benefit is most evident on hardware with fewer cores.
This is the result on an 4-core laptop:
||before:|after:
|---|---|---|
|large_mesh_editing:|Average: 5.203638 FPS|Average: 5.398925 FPS
||rdata 15ms iter 43ms (frame 193ms)|rdata 14ms iter 36ms (frame 187ms)
Differential Revision: https://developer.blender.org/D11558
This is an adaptation of {D11488}.
A disadvantage of manually setting the iter ranges per thread is that
we don't know how many threads are running in the background and so we
don't know how to best distribute the ranges.
To solve this limitation we can use `parallel_reduce` and thus let the
driver choose the best distribution of ranges among the threads.
This proved to be especially beneficial for computers with few cores.
**Benchmarking:**
Here's the result on an 4-core laptop:
||master:|PATCH:
|---|---|---|
|large_mesh_editing:|Average: 5.203638 FPS|Average: 5.398925 FPS
||rdata 15ms iter 43ms (frame 193ms)|rdata 14ms iter 36ms (frame 187ms)
Here's the result on an 8-core PC:
||master:|PATCH:
|---|---|---|
|large_mesh_editing:|Average: 15.267482 FPS|Average: 15.906881 FPS
||rdata 9ms iter 28ms (frame 65ms)|rdata 9ms iter 25ms (frame 63ms)
|large_mesh_editing_ledge: |Average: 15.145966 FPS|Average: 15.520474 FPS
||rdata 9ms iter 29ms (frame 65ms)|rdata 9ms iter 25ms (frame 64ms)
|looptris_test:|Average: 4.001917 FPS|Average: 4.061105 FPS
||rdata 12ms iter 90ms (frame 236ms)|rdata 12ms iter 87ms (frame 230ms)
|subdiv_mesh_cage_and_final:|Average: 1.917769 FPS|Average: 1.971790 FPS
||rdata 7ms iter 37ms (frame 261ms)|rdata 7ms iter 31ms (frame 258ms)
||rdata 7ms iter 38ms (frame 252ms)|rdata 7ms iter 33ms (frame 249ms)
|subdiv_mesh_final_only:|Average: 6.387240 FPS|Average: 6.591251 FPS
||rdata 3ms iter 25ms (frame 151ms)|rdata 3ms iter 16ms (frame 145ms)
|subdiv_mesh_final_only_ledge:|Average: 6.247393 FPS|Average: 6.596024 FPS
||rdata 3ms iter 26ms (frame 158ms)|rdata 3ms iter 16ms (frame 148ms)
**Notes:**
- The improvement can only be noticed if all extracts are multithreaded.
- This patch touches different areas of the code, so it can be split into another patch if the idea is accepted.
These screenshots show how threads behave in a quadcore:
Master:
{F10164664}
Patch:
{F10164666}
Differential Revision: https://developer.blender.org/D11558
ffmpeg_generic_seek_workaround did work properly and our start pts
calculation was wrong.
Reviewed By: Richard Antalik
Differential Revision: http://developer.blender.org/D11562
Because of the added sanity checks in rB14508ef100c9 (D11492), seeking
in proxies would not work correctly any more. This is because it wasn't
working as intended before, but in most cases this wouldn't be
noticeable. However now when the sanity checks are tripped it is very
noticeable that something is wrong
The indexer tried to use dts values for time stamps when we used pts in
our decode functions to get the time positions. This would make it
start in the wrong GOP frames when searching. Now that we enforce no
crossing of GOP frames when decoding after seek, this would lead to
issues.
Now we correctly use pts (or dts if pts is not available) and thus we
don't have any seeking issues because of time stamp format missmatch.
Reviewed By: Richard Antalik
Differential Revision: http://developer.blender.org/D11561
When sampling ImBuf can be a char or a float buffer. Current sampling
functions added overhead by checking which kind of buffer was passed
every pixel that was sampled. When performing image processing this
check can be removed outside the inner loop adding 5% of performance
increase in the `IMB_transform` operator.
Inside the sequencer the cropping and transform of images/buffers were
implemented locally. This reduced the optimizations that a compiler
could do and added confusing code styles. This patch adds
`IMB_transform` to reduce the confusion and increases compiler
optimizations as more code can be inlined and we can keep track of
indices inside the inner loop.
This increases end-user performance by 30% when playing back aa video
in VSE.
Reviewed By: ISS, zeddb
Differential Revision: https://developer.blender.org/D11549
This patch changes occurrences of percentage to factor.
There are some usages of percentage left in there on purpose.
They are distinguished as follows:
- factor is 0-1 float
- percentage is 0-100 int
Ref D11361
Reviewed by: sybren, campbellbarton
Prepare node for conversion to Geometry Nodes.
There should be no functional changes.
Reviewed By: HooglyBoogly
Differential Revision: https://developer.blender.org/D11506
When using non-default system separator in filename path, code would end up
with an absolute path mixing regular and alternative separator,
confusing the rest of the path manipulations later on.
So this commit add proper replacements of alternative separators, and
path normalization.
Splitting out thread safe iteration logic means regular iteration
isn't checking for the thread-safe pointer each step.
This gives a small but measurable overall performance gain of 2-3%
when redrawing a high-poly mesh.
Ref D11564
Reviewed By: mont29
This change was prompted by D6408 which moves thumbnail extraction into
a shared function that happens use these endian defines but only links
blenlib.
There is no need for these defines to be associated with globals
so move into their own header.
Allows to define properties which will have proper units displayed
in the interface. The internal storage is expected to be seconds
(which matches how other times are stored in Blender).
Is not immediately used in Blender, but is required for the upcoming
feature in Cycles X (D11526)
The naming does not sound very exciting, but can't think of anything
better either.
For test it probably easiest to define FloatProperty with subdtype
of TIME_ABSOLUTE.
Differential Revision: https://developer.blender.org/D11532
Previously, a vertex from destination mesh would always be added to all
transferred vgroup (with a 0.0 weight), even if none of its matching
sources in source mesh belonged to the matching source vgroups.
Now a destination vertex is only added to a given destination vgroup if
at least one of its source vertices belong to the matching source
vgroup.
Issue found and initial investigation by @pls in D11524, thanks!
TBBmalloc_proxy already takes care of any allocations
being done from MSVC compiled code, some of the dependencies
like GMP cannot be build with MSVC and we have to use
mingw to build them. mingw however links against the older
msvcrt.dll for its allocation needs, which TBBMallocProxy
does not hook.
GMP has an option to supply your own allocation functions
so we can still manually redirect them to TBBMalloc.
In a test-file with a boolean geometry node, this patch
uses 32s effective CPU time compared to 52s before.
Differential Revision: https://developer.blender.org/D11435
Reviewed by Campbell Barton, Ray Molenkamp
Remove an unnecessary call to pose_slide_mouse_update_percentage
That call was overriding the typed value
Reviewed By: #animation_rigging, Sybren A. Stüvel
Differential Revision: https://developer.blender.org/D11395
Ref D11395
To draw the overlay stats in columns the strings must be measured to
find the longest one. In some circumstances this measurement can be
incorrect. We draw the text with a specific size yet do not explicitly
set the size before calling BLF_size. This patch properly sets the size
so that the measurement will match what will be used for output.
See T88799 for examples of measurement failure.
Differential Revision: https://developer.blender.org/D11504
Reviewed by Hans Goudey
First, expand on the interpolation to evaluated points with a templated
helper function, and a function that takes a GSPan. Next, add a set of
functions to `Spline` for interpolating at arbitrary intervals between
the evaluated points. The code for doing that isn't that complicated
anyway, but it's nice to avoid repeating, and it might make it easier
to unroll the special cases for the first and last points if we require
the index factors to be sorted.
This commit adds a node to output the convex hull of any input geometry
as a mesh, which is an enclosing geometry around a set of points.
All geometry types are supported, besides volumes.
The code supports operating on instances to avoid copying all input
geometry before the operation. The implementation uses the same backend
as the operation in edit mode, but uses Mesh directly instead of BMesh.
Attribute transfer is not supported currently, but would be a point of
improvement for the future if it can work in a predictable way on
different geometry input types.
Differential Revision: https://developer.blender.org/D10925
Note that these changes are limited simple cases as these kinds of
changes could allow for errors when refactoring code when the known
state is not so obvious.
