Previously, `VArrayImpl` had a `materialize` and `materialize_to_uninitialized`
function. Now both are merged into one with an additional `bool
dst_is_uninitialized` parameter. The same is done for the
`materialize_compressed` method as all as `GVArrayImpl`.
While this kind of merging is typically not ideal, it reduces the binary size by
~200kb while being basically free performance wise. The cost of this predictable
boolean check is expected to be negligible even if only very few indices are
materialized. Additionally, in most cases, this parameter does not even have to
be checked, because for trivial types it does not matter if the destination
array is already initialized or not when overwriting it.
It saves this much memory, because there are quite a few implementations being
generated with e.g. `VArray::from_func` and a lot of code was duplicated for
each instantiation.
This changes only the actual `(G)VArrayImpl`, but not the `VArray` and `GVArray`
API which is typically used to work with virtual arrays.
Pull Request: https://projects.blender.org/blender/blender/pulls/145144
This patch turns the options of the Color Balance node into inputs.
In the process, each of the wheels were split into two inputs, a base
float and a color. For instance, Gain is controlled using both a Base
Gain and Color Gain, the former controls the gain for all channels while
the latter controls it per channel.
Reference #137223.
Pull Request: https://projects.blender.org/blender/blender/pulls/138610
This patch turns the options of the Chroma Key node into inputs.
In the process, the minimum and maximum angles were renamed to Minimum
and Maximum for consistency with other matte nodes.
Reference #137223.
Pull Request: https://projects.blender.org/blender/blender/pulls/137812
This patch implements the multi-function procedure operation for the new
CPU compositor, which is a concrete implementation of the PixelOperation
abstraction, much like ShaderOperation, but uses the FN system to more
efficiently evaluate a group of pixel-wise operations.
A few changes were done to FN to support development. The multi-function
builder now allows retrieving the built function. A new builder method
construct_and_set_matching_fn_cb was added to allow using the SI_SO
builders with non static functions. A few other SI_SO were added to. And
a CPP type for float4 was added.
Additionally, the Gamma, Math, Brightness, and Normal nodes were
implemented as an example. The Math node implementation reused the
existing GN math node implementation, so the code was moved to a common
file.
Reference #125968.
Pull Request: https://projects.blender.org/blender/blender/pulls/126988
Along with the 4.1 libraries upgrade, we are bumping the clang-format
version from 8-12 to 17. This affects quite a few files.
If not already the case, you may consider pointing your IDE to the
clang-format binary bundled with the Blender precompiled libraries.
Listing the "Blender Foundation" as copyright holder implied the Blender
Foundation holds copyright to files which may include work from many
developers.
While keeping copyright on headers makes sense for isolated libraries,
Blender's own code may be refactored or moved between files in a way
that makes the per file copyright holders less meaningful.
Copyright references to the "Blender Foundation" have been replaced with
"Blender Authors", with the exception of `./extern/` since these this
contains libraries which are more isolated, any changed to license
headers there can be handled on a case-by-case basis.
Some directories in `./intern/` have also been excluded:
- `./intern/cycles/` it's own `AUTHORS` file is planned.
- `./intern/opensubdiv/`.
An "AUTHORS" file has been added, using the chromium projects authors
file as a template.
Design task: #110784
Ref !110783.
A lot of files were missing copyright field in the header and
the Blender Foundation contributed to them in a sense of bug
fixing and general maintenance.
This change makes it explicit that those files are at least
partially copyrighted by the Blender Foundation.
Note that this does not make it so the Blender Foundation is
the only holder of the copyright in those files, and developers
who do not have a signed contract with the foundation still
hold the copyright as well.
Another aspect of this change is using SPDX format for the
header. We already used it for the license specification,
and now we state it for the copyright as well, following the
FAQ:
https://reuse.software/faq/
Goals of this refactor:
* Reduce memory consumption of `IndexMask`. The old `IndexMask` uses an
`int64_t` for each index which is more than necessary in pretty much all
practical cases currently. Using `int32_t` might still become limiting
in the future in case we use this to index e.g. byte buffers larger than
a few gigabytes. We also don't want to template `IndexMask`, because
that would cause a split in the "ecosystem", or everything would have to
be implemented twice or templated.
* Allow for more multi-threading. The old `IndexMask` contains a single
array. This is generally good but has the problem that it is hard to fill
from multiple-threads when the final size is not known from the beginning.
This is commonly the case when e.g. converting an array of bool to an
index mask. Currently, this kind of code only runs on a single thread.
* Allow for efficient set operations like join, intersect and difference.
It should be possible to multi-thread those operations.
* It should be possible to iterate over an `IndexMask` very efficiently.
The most important part of that is to avoid all memory access when iterating
over continuous ranges. For some core nodes (e.g. math nodes), we generate
optimized code for the cases of irregular index masks and simple index ranges.
To achieve these goals, a few compromises had to made:
* Slicing of the mask (at specific indices) and random element access is
`O(log #indices)` now, but with a low constant factor. It should be possible
to split a mask into n approximately equally sized parts in `O(n)` though,
making the time per split `O(1)`.
* Using range-based for loops does not work well when iterating over a nested
data structure like the new `IndexMask`. Therefor, `foreach_*` functions with
callbacks have to be used. To avoid extra code complexity at the call site,
the `foreach_*` methods support multi-threading out of the box.
The new data structure splits an `IndexMask` into an arbitrary number of ordered
`IndexMaskSegment`. Each segment can contain at most `2^14 = 16384` indices. The
indices within a segment are stored as `int16_t`. Each segment has an additional
`int64_t` offset which allows storing arbitrary `int64_t` indices. This approach
has the main benefits that segments can be processed/constructed individually on
multiple threads without a serial bottleneck. Also it reduces the memory
requirements significantly.
For more details see comments in `BLI_index_mask.hh`.
I did a few tests to verify that the data structure generally improves
performance and does not cause regressions:
* Our field evaluation benchmarks take about as much as before. This is to be
expected because we already made sure that e.g. add node evaluation is
vectorized. The important thing here is to check that changes to the way we
iterate over the indices still allows for auto-vectorization.
* Memory usage by a mask is about 1/4 of what it was before in the average case.
That's mainly caused by the switch from `int64_t` to `int16_t` for indices.
In the worst case, the memory requirements can be larger when there are many
indices that are very far away. However, when they are far away from each other,
that indicates that there aren't many indices in total. In common cases, memory
usage can be way lower than 1/4 of before, because sub-ranges use static memory.
* For some more specific numbers I benchmarked `IndexMask::from_bools` in
`index_mask_from_selection` on 10.000.000 elements at various probabilities for
`true` at every index:
```
Probability Old New
0 4.6 ms 0.8 ms
0.001 5.1 ms 1.3 ms
0.2 8.4 ms 1.8 ms
0.5 15.3 ms 3.0 ms
0.8 20.1 ms 3.0 ms
0.999 25.1 ms 1.7 ms
1 13.5 ms 1.1 ms
```
Pull Request: https://projects.blender.org/blender/blender/pulls/104629
This allows auto-vectorization to happen when the a multi-function is
evaluated in "materialized" mode, i.e. it is processed in chunks where
all input and outputs values are stored in contiguous arrays.
It also unifies the handling input, mutable and output parameters a bit.
Now they all can use tempory buffers in the same way.
This simplifies the code enough so that msvc is able to unroll and
vectorize some multi-functions like simple addition.
The performance improvements are almost as good as the GCC
improvements shown in D16942 (for add and multiply at least).
This mainly helps GCC catch up with Clang in terms of field evaluation
performance in some cases. In some cases this patch can speedup
field evaluation 2-3x (e.g. when there are many float math nodes).
See D16942 for a more detailed benchmark.
This moves all multi-function related code in the `functions` module
into a new `multi_function` namespace. This is similar to how there
is a `lazy_function` namespace.
The main benefit of this is that many types names that were prefixed
with `MF` (for "multi function") can be simplified.
There is also a common shorthand for the `multi_function` namespace: `mf`.
This is also similar to lazy-functions where the shortened namespace
is called `lf`.
This avoids a move of the signature after building it. Tthe value had
to be moved out of `MFSignatureBuilder` in the `build` method.
This also makes the naming a bit less confusing where sometimes
both the `MFSignature` and `MFSignatureBuilder` were referred
to as "signature".
* New `build_mf` namespace for the multi-function builders.
* The type name of the created multi-functions is now "private",
i.e. the caller has to use `auto`. This has the benefit that the
implementation can change more freely without affecting
the caller.
* `CustomMF` does not use `std::function` internally anymore.
This reduces some overhead during code generation and at
run-time.
* `CustomMF` now supports single-mutable parameters.
This refactors how devirtualization is done in general and how
multi-functions use it.
* The old `Devirtualizer` class has been removed in favor of a simpler
solution. It is also more general in the sense that it is not coupled
with `IndexMask` and `VArray`. Instead there is a function that has
inputs which control how different types are devirtualized. The
new implementation is currently less general with regard to the number
of parameters it supports. This can be changed in the future, but
does not seem necessary now and would make the code less obvious.
* Devirtualizers for different types are now defined in their respective
headers.
* The multi-function builder works with the `GVArray` stored in `MFParams`
directly now, instead of first converting it to a `VArray<T>`. This reduces
some constant overhead, which makes the multi-function slightly
faster. This is only noticable when very few elements are processed though.
No functional changes or performance regressions are expected.
a5e7657cee didn't account for slices of zero sizes, and the asserts
were slightly incorrect otherwise. Also, the change didn't apply to
`Span`, only `MutableSpan`, which was a mistake. This also adds "safe"
methods to `IndexMask`, and switches function calls where necessary.
This is the conventional way of dealing with unused arguments in C++,
since it works on all compilers.
Regex find and replace: `UNUSED\((\w+)\)` -> `/*$1*/`
Goals:
* Better high level control over where devirtualization occurs. There is always
a trade-off between performance and compile-time/binary-size.
* Simplify using array devirtualization.
* Better performance for cases where devirtualization wasn't used before.
Many geometry nodes accept fields as inputs. Internally, that means that the
execution functions have to accept so called "virtual arrays" as inputs. Those
can be e.g. actual arrays, just single values, or lazily computed arrays.
Due to these different possible virtual arrays implementations, access to
individual elements is slower than it would be if everything was just a normal
array (access does through a virtual function call). For more complex execution
functions, this overhead does not matter, but for small functions (like a simple
addition) it very much does. The virtual function call also prevents the compiler
from doing some optimizations (e.g. loop unrolling and inserting simd instructions).
The solution is to "devirtualize" the virtual arrays for small functions where the
overhead is measurable. Essentially, the function is generated many times with
different array types as input. Then there is a run-time dispatch that calls the
best implementation. We have been doing devirtualization in e.g. math nodes
for a long time already. This patch just generalizes the concept and makes it
easier to control. It also makes it easier to investigate the different trade-offs
when it comes to devirtualization.
Nodes that we've optimized using devirtualization before didn't get a speedup.
However, a couple of nodes are using devirtualization now, that didn't before.
Those got a 2-4x speedup in common cases.
* Map Range
* Random Value
* Switch
* Combine XYZ
Differential Revision: https://developer.blender.org/D14628
- Missing star prefix.
- Unnecessary indentation.
- Blank line after dot-points
(otherwise doxygen merges with the previous dot-point).
- Use back-slash for doxygen commands.
- Correct spelling.
This implements two optimizations:
* Reduce virtual function call overhead when a non-standard virtual
array is used as input.
* Use a lambda in `type_conversion.cc`.
In my test setup, which creates a float attribute filled with the index,
the running time drops from `4.0 ms` to `2.0 ms`.
Differential Revision: https://developer.blender.org/D14585
This extracts the inner loops into a separate function.
There are two main reasons for this:
* Allows using `__restrict` to indicate that no other parameter
aliases with the output array. This allows for better optimization.
* Makes it easier to search for the generated assembly code,
especially with the `BLI_NOINLINE`.
This simplifies debugging, and can help improve performance
by making it easier for the compiler.
More optimization might still be possible by using `__restrict` in
a few places.
Use a shorter/simpler license convention, stops the header taking so
much space.
Follow the SPDX license specification: https://spdx.org/licenses
- C/C++/objc/objc++
- Python
- Shell Scripts
- CMake, GNUmakefile
While most of the source tree has been included
- `./extern/` was left out.
- `./intern/cycles` & `./intern/atomic` are also excluded because they
use different header conventions.
doc/license/SPDX-license-identifiers.txt has been added to list SPDX all
used identifiers.
See P2788 for the script that automated these edits.
Reviewed By: brecht, mont29, sergey
Ref D14069
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.
Since fields were committed to master, socket inspection did
not work correctly for all socket types anymore. Now the same
functionality as before is back. Furthermore, fields that depend
on some input will now show the inputs in the socket inspection.
I added support for evaluating constant fields more immediately.
This has the benefit that the same constant field is not evaluated
more than once. It also helps with making the field independent
of the multi-functions that it uses. We might still want to change
the ownership handling for the multi-functions of nodes a bit,
but that can be done separately.
Differential Revision: https://developer.blender.org/D12444
This implements the initial core framework for fields and anonymous
attributes (also see T91274).
The new functionality is hidden behind the "Geometry Nodes Fields"
feature flag. When enabled in the user preferences, the following
new nodes become available: `Position`, `Index`, `Normal`,
`Set Position` and `Attribute Capture`.
Socket inspection has not been updated to work with fields yet.
Besides these changes at the user level, this patch contains the
ground work for:
* building and evaluating fields at run-time (`FN_fields.hh`) and
* creating and accessing anonymous attributes on geometry
(`BKE_anonymous_attribute.h`).
For evaluating fields we use a new so called multi-function procedure
(`FN_multi_function_procedure.hh`). It allows composing multi-functions
in arbitrary ways and supports efficient evaluation as is required by
fields. See `FN_multi_function_procedure.hh` for more details on how
this evaluation mechanism can be used.
A new `AttributeIDRef` has been added which allows handling named
and anonymous attributes in the same way in many places.
Hans and I worked on this patch together.
Differential Revision: https://developer.blender.org/D12414
* Reduce code duplication.
* Give methods more standardized names (e.g. `move_to_initialized` -> `move_assign`).
* Support wrapping arbitrary C++ types, even those that e.g. are not copyable.
In some multi-functions (such as a simple add function), the virtual method
call overhead to access array elements adds significant overhead. For these
simple functions it makes sense to generate optimized versions for different
types of virtual arrays. This is done by giving the compiler all the information
it needs to devirtualize virtual arrays.
In my benchmark this speeds up processing a lot of data with small function 2-3x.
This devirtualization should not be done for larger functions, because it increases
compile time and binary size, while providing a negilible performance benefit.
Previously, the signature of a `MultiFunction` was always embedded into the function.
There are two issues with that. First, `MFSignature` is relatively large, because it contains
multiple strings and vectors. Secondly, constructing it can add overhead that should not
be necessary, because often the same signature can be reused.
The solution is to only keep a pointer to a signature in `MultiFunction` that is set during
construction. Child classes are responsible for making sure that the signature lives
long enough. In most cases, the signature is either embedded into the child class or
it is allocated statically (and is only created once).
