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
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
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
File output node always received the resolution from the first socket.
When that socket didn't had a link it would use a resolution of 0,0.
What lead to not saving the file at all.
This only effected Multi layer OpenEXR files.
This change would go over all the links to find the first valid
resolution.
For debugging purposes to convert the internal state of the
NodeOperationBuilder to a graphviz.
Usage:
std::cout << *this << "\n";
Inside any method of the NodeOperationBuilder.
SocketReader was added as an easier to understand interface on top of
the NodeOperation. It was implemented as a base class of the
NodeOperation and adds an additional hierarchy level.
Ths change replaces the abstract class with a typedef. In the end we
want to remove the typedef but will wait for some new nodes before doing
so.
This replaces header include guards with `#pragma once`.
A couple of include guards are not removed yet (e.g. `__RNA_TYPES_H__`),
because they are used in other places.
This patch has been generated by P1561 followed by `make format`.
Differential Revision: https://developer.blender.org/D8466
Surrounding includes with an 'extern "C"' block is not necessary anymore.
Also that made it harder to add any C++ code to some headers, or include headers
that have "optional" C++ code like `MEM_guardedalloc.h`.
I tested compilation on linux and windows (and got help from @LazyDodo).
If this still breaks compilation due to some linker error, the header containing
the symbol in question is probably missing an 'extern "C"' block.
Differential Revision: https://developer.blender.org/D7653
BF-admins agree to remove header information that isn't useful,
to reduce noise.
- BEGIN/END license blocks
Developers should add non license comments as separate comment blocks.
No need for separator text.
- Contributors
This is often invalid, outdated or misleading
especially when splitting files.
It's more useful to git-blame to find out who has developed the code.
See P901 for script to perform these edits.
For now it was mainly about OpenCL wrangler being duplicated
between Cycles and Compositor, but with OpenSubdiv work those
wranglers were gonna to be duplicated just once again.
This commit makes it so Cycles and Compositor uses wranglers
from this repositories:
- https://github.com/CudaWrangler/cuew
- https://github.com/OpenCLWrangler/clew
This repositories are based on the wranglers we used before
and they'll be likely continued maintaining by us plus some
more players in the market.
Pretty much straightforward change with some tricks in the
CMake/SCons to make this libs being passed to the linker
after all other libraries in order to make OpenSubdiv linked
against those wranglers in the future.
For those who're worrying about Cycles being less standalone,
it's not truth, it's rather more flexible now and in the future
different wranglers might be used in Cycles. For now it'll
just mean those libs would need to be put into Cycles repository
together with some other libs from Blender such as mikkspace.
This is mainly platform maintenance commit, should not be any
changes to the user space.
Reviewers: juicyfruit, dingto, campbellbarton
Reviewed By: juicyfruit, dingto, campbellbarton
Differential Revision: https://developer.blender.org/D707
Proxy operations from muted nodes would still create conversion
operations where the datatypes don't match, which creates unexpected
behavior. Arguably datatype conversion could still happen even when the
main operation is muted, but this would be a design change and so is
disabled now.
Many parts of the compositor are unnecessarily complicated. This patch
aims at reducing the complexity of writing nodes and making the code
more transparent.
== Separating Nodes and Operations ==
Currently these are both mixed in the same graph, even though they have
very different purposes and are used at distinct stages in the
compositing process. The patch introduces dedicated graph classes for
nodes and for operations.
This removes the need for a lot of special case checks (isOperation etc.)
and explicit type casts. It simplifies the code since it becomes clear
at every stage what type of node we are dealing with. The compiler can
use static typing to avoid common bugs from mixing up these types and
fewer runtime sanity checks are needed.
== Simplified Node Conversion ==
Converting nodes to operations was previously based on "relinking", i.e.
nodes would start with by mirroring links in the Blender DNA node trees,
then add operations and redirect these links to them. This was very hard
to follow in many cases and required a lot of attention to avoid invalid
states.
Now there is a helper class called the NodeConverter, which is passed to
nodes and implements a much simpler API for this process. Nodes can add
operations and explicit connections as before, but defining "external"
links to the inputs/outputs of the original node now uses mapping
instead of directly modifying link data. Input data (node graph) and
result (operations graph) are cleanly separated.
== Removed Redundant Data Structures ==
A few redundant data structures have been removed, notably the
SocketConnection. These are only needed temporarily during graph
construction. For executing the compositor operations it is perfectly
sufficient to store only the direct input link pointers. A common
pointer indirection is avoided this way (which might also give a little
performance improvement).
== Avoid virtual recursive functions ==
Recursive virtual functions are evil. They are very hard to follow
during debugging. At least in the parts this patch is concerned with
these functions have been replaced by a non-virtual recursive core
function (which might then call virtual non-recursive functions if
needed). See for example NodeOperationBuilder::group_operations.
Debug code for graphviz output moved to a dedicated file COM_Debug.h/cpp.
The DebugInfo class has only static functions, which are called from a number of places to keep track of what is happening in the compositor. If debugging is disabled these are just inline stubs, so we
don't need #ifdefs everywhere and don't get any overhead.
The graphviz output is much more useful now. DebugInfo keeps track of node names in a static string map for meaningful names. It uses a number of colors for various special operation classes.
ExecutionGroups are indicated in graphviz with clusters.
Currently the graphviz .dot files are stored in the BLI_temporary_dir() folder. A separate dot file is generated for each stage of the ExecutionGroup scheduling, this is intended to give some idea of the
compositor progress, but could still be improved.
Issue was caused by file output node actually,
The thing here is, compositor output does have fixed
resolution and we could predict how to map coordinates
for border and cropping in that case.
But viewers and file output nodes are currently totally
depending on an input resolution. Could not see how
border could be applied reliably in this cases.
Disabling border option for file output node, so
now it shall behave the same way as it was before.
Discovered issues when using cropping to render border,
namely there's an offset in viewer nodes and previews,
but this is separate issue i guess (file output seems
to work fine). Will revisit this issue in next days.
