Since the introduction of storage buffers in Blender, the calling
code has been responsible for ensuring the buffer meets allocation
requirements. All backends require the allocation size to be divisible
by 16 bytes. Until now, this was sufficient, but with GPU subdivision
changes, an external library must also adhere to these requirements.
For OpenSubdiv (OSD), some buffers are not 16-byte aligned, leading
to potential misallocation. Currently, this is mitigated by allocating
a few extra bytes, but this approach has the drawback of potentially
reading unintended bytes beyond the source buffer.
This PR adopts a similar approach to vertex buffers: the backend handles
extra byte allocation while ensuring data uploads and downloads function
correctly without requiring those additional bytes.
No changes were needed for Metal, as its allocation size is already
aligned to 256 bytes.
**Alternative solutions considered**:
- Copying the CPU buffer to a larger buffer when needed (performance impact).
- Modifying OSD buffers to allocate extra space (requires changes to an external library).
- Implementing GPU_storagebuf_update_sub.
Ref #135873
Pull Request: https://projects.blender.org/blender/blender/pulls/135716
2nd derivatives aren't used within Blender but all
logic branches end up with checking for it. This PR
removes the 2nd derivatives from the code so we
can skip this test.
Pull Request: https://projects.blender.org/blender/blender/pulls/135745
Blender already had its own copy of OpenSubDiv containing some local fixes
and code-style. This code still used gl-calls. This PR updates the calls
to use GPU module. This allows us to use OpenSubDiv to be usable on other
backends as well.
This PR was tested on OpenGL, Vulkan and Metal. Metal can be enabled,
but Vulkan requires some API changes to work with loose geometry.
# Considerations
**ShaderCreateInfo**
intern/opensubdiv now requires access to GPU module. This to create buffers
in the correct context and trigger correct dispatches. ShaderCreateInfo is used
to construct the shader for cross compilation to Metal/Vulkan. However opensubdiv
shader caching structures are still used.
**Vertex buffers vs storage buffers**
Implementation tries to keep as close to the original OSD implementation. If
they used storage buffers for data, we will use GPUStorageBuf. If it uses vertex
buffers, we will use gpu::VertBuf.
**Evaluator const**
The evaluator cannot be const anymore as the GPU module API only allows
updating SSBOs when constructing. API could be improved to support updating
SSBOs.
Current implementation has a change to use reads out of bounds when constructing
SSBOs. An API change is in the planning to remove this issue. This will be fixed in
an upcoming PR. We wanted to land this PR as the visibility of the issue is not
common and multiple other changes rely on this PR to land.
Pull Request: https://projects.blender.org/blender/blender/pulls/135296
`OpenSubdiv_Buffer` is a wrapper that was introduced at the time
that Blender couldn't use CPP directly. It contains a pointer to
a VertBuf and callbacks to use GPU module on that buffer.
This PR replaces OpenSubdiv_Buffer with `blender::gpu::VertBuf` and
removes the wrapper.
NOTE: OpenSubdiv tests are added to blender_test executable to make the
library dependencies not to complicated.
Pull Request: https://projects.blender.org/blender/blender/pulls/135389
Similar to 5e46e3d28a.
This commit replaces the C-API version of `OpenSubdiv_Evaluator`
with direct calls to `EvalOutputAPI`. This removes a level of indirection,
theoretically reducing function call overhead, but also making the whole
system easier to understand and easier to modify. The downside is
further spread of `WITH_OPENSUBDIV` into the code, but I think that
can be improved in the future relatively easily once more of this sort
of change is finished.
Pull Request: https://projects.blender.org/blender/blender/pulls/128278
Remove the indirection previously used for the topology refiner
to separate C and C++ code. Instead retrieve the base level in
calling code and call opensubdiv API functions directly. This
avoids copying arrays of mesh indices and should reduce
function call overhead since index retrieval can now be inlined.
It also lets us remove a lot of boilerplate shim code.
The downside is increased need for WITH_OPENSUBDIV defines
in various parts of blenkernel, but I think that is required to avoid
the previous indirection and have the kernel deal with OpenSubdiv
more directly.
Pull Request: https://projects.blender.org/blender/blender/pulls/120825
Replace the C-class pattern function pointers with actual class methods.
Other than the obvious benefit of not requiring the "this" pointer to be
explicitly passed into every function call, this will make it much simpler
to remove the entire C-API class and replace it with its "impl" next.
For that next step we need to expose code to the implementation
of the topology refiner, so instead of defining stubs locally in the
opensubdiv intern class, we spread some WITH_OPENSUBDIV checks
in the blenkernel. As far as I know this is the only way to remove the
intermediate C-API and call opensubdiv functions directly from there.
Avoid unnecessary indirect includes for the module, which would slow
down compilation. Avoid "using std" which is generally not considered
good practice since it's helpful to see what namespace things are from.
This is just a general cleanup of the area. I was trying to understand
it better to reduce redundant mesh topology storage but finding that
some intermediate cleanups would be helpful to ease the change.
Pull Request: https://projects.blender.org/blender/blender/pulls/120743
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.
All the relevant code is C++ now, so we don't need to complicate things
with the trip through C anymore. We will still need some wrappers, since
opensubdiv is an optional dependency though. The goal is to make it
simpler to remove the unnecessary/costly abstraction levels between
Blender mesh data and the opensubdiv code.
The goal is to solve confusion of the "All rights reserved" for licensing
code under an open-source license.
The phrase "All rights reserved" comes from a historical convention that
required this phrase for the copyright protection to apply. This convention
is no longer relevant.
However, even though the phrase has no meaning in establishing the copyright
it has not lost meaning in terms of licensing.
This change makes it so code under the Blender Foundation copyright does
not use "all rights reserved". This is also how the GPL license itself
states how to apply it to the source code:
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) <year> <name of author>
This program is free software ...
This change does not change copyright notice in cases when the copyright
is dual (BF and an author), or just an author of the code. It also does
mot change copyright which is inherited from NaN Holding BV as it needs
some further investigation about what is the proper way to handle it.
For example
```
OIIOOutputDriver::~OIIOOutputDriver()
{
}
```
becomes
```
OIIOOutputDriver::~OIIOOutputDriver() {}
```
Saves quite some vertical space, which is especially handy for
constructors.
Pull Request: https://projects.blender.org/blender/blender/pulls/105594
With libepoxy we can choose between EGL and GLX at runtime, as well as
dynamically open EGL and GLX libraries without linking to them.
This will make it possible to build with Wayland, EGL, GLVND support while
still running on systems that only have X11, GLX and libGL. It also paves
the way for headless rendering through EGL.
libepoxy is a new library dependency, and is included in the precompiled
libraries. GLEW is no longer a dependency, and WITH_SYSTEM_GLEW was removed.
Includes contributions by Brecht Van Lommel, Ray Molenkamp, Campbell Barton
and Sergey Sharybin.
Ref T76428
Differential Revision: https://developer.blender.org/D15291
openSubdiv_init() would detect available evaluators before any OpenGL context
exists, causing a crash with libepoxy. This test however is redundant as we
already check the requirements on the Blender side through the GPU API.
To simplify things, completely remove the device detection in the opensubdiv
module and reduce the evaluators to just CPU and GPU. The plan here is to move
to the GPU module abstraction over OpenGL/Metal/Vulkan and so all these
different backends no longer make sense.
This also removes the user preference for OpenSubdiv compute device, which was
not used for the new GPU subdivision implementation.
Ref D15291
Differential Revision: https://developer.blender.org/D15470
When multiple objects are in edit mode, UVs for the objects, except for
the first one (in rendering order) appear corrupted. The corruption is
because the UVs are not evaluated as the compute shader is not bound,
thus we read unitialized memory.
We keep track of the currently bound shader in the GPU context in order
to avoid unnecessary shader switches in case the same shader is used in
consecutive calls. However, the shader used by the OpenSubdiv evaluator
is not part of Blender and therefore not tracked via the GPU context.
When extracting UVs for multiple objects, we only ever run a single
shader (FVar evaluation). However, between the compute calls, we also
call the OpenSubdiv stencil evaluation shader, which uses `glUseProgram`
modifying the current program, outside of our control, which then also
unbinds the Blender compute shader making the compute dispatch fail ("No
active compute shader").
The fact that extracting the UVs for the first rendered object works is
because another (Blender) shader was bound in the GPU context prior to
our binding of our evaluation shader.
To fix this, we remember, in the OpenSubdiv evaluator, the current
program so that it can be reset after the stencil program is done.
Differential Revision: https://developer.blender.org/D15064
Subdivision did not properly update when evaluating first without and then with
orco coordinates. Now update the subdivision evaluator settings every time, and
reallocate the vertex data buffer when needed.
there is an additional issue in this file where orco coordinates are not
available immediately on the first frame when they should be, and only appear
on the second frame. However that is an old limitation related to the depsgraph
not getting re-evaluated on viewport display mode changes, here we just fix the
crash.
This uses the recently introduced evaluator's vertex
data to smoothly interpolate original coordinates instead
of using linear interpolation.
The orcos are interpolated at the same time as positions
and as such, the specific subdivision routine for the
orco extractor has been removed. The patch evaluation
shader uses a definition to enable code specific to
orco evaluation.
Since the orco layer may not have been requested on first
render, and since orco data is now stored in the OpenSubDiv
evaluator, the evaluator needs to be recreated if an
orco layer is suddenly available. For this, a callback
to check if the evaluator has the data was added. This is
added to the evaluator as the `Subdiv` cache stored in the
modifier is invalidated less often than the Mesh batch cache
and so leads to fewer evaluator recreations.
Differential Revision: https://developer.blender.org/D14999
This makes changes to the opensubdiv module to support additional vertex data
besides the vertex position, that is smootly interpolated the same way. This is
different than varying data which is interpolated linearly.
Fixes T96596: wrong generated texture coordinates with GPU subdivision. In that
bug lazy subdivision would not interpolate orcos.
Later on, this implementation can also be used to remove the modifier stack
mechanism where modifiers are evaluated a second time for orcos, which is messy
and inefficient. But that's a more risky change, this is just the part to fix
the bug in 3.2.
Differential Revision: https://developer.blender.org/D14973
This was caused by the use of a reserved keyword macro that is not
directly used but causes an error on some compiler.
Change the occurences to not match the macros.
Coarse meshes with high polycount would show as corrupted when GPU
subdivision is used with AMD cards This was caused by the OpenSubdiv
library not taking `GL_MAX_COMPUTE_WORK_GROUP_COUNT` into account when
dispatching computes. AMD drivers tend to set the limit lower than
NVidia ones (2^16 for the former, and 2^32 for the latter, at least
on my machine).
This moves the `GLComputeEvaluator` from the OpenSubdiv library into
`intern/opensubdiv` and modifies it to compute a dispatch size in a
similar way as for the draw code: we split the dispatch size into a 2
dimensional value based on `GL_MAX_COMPUTE_WORK_GROUP_COUNT` and
manually compute an index in the shader.
We could have patched the OpenSubdiv library and sent the fix upstream
(which can still be done), however, moving it to our side allows us to
better control the `GLComputeEvaluator` and in the future remove some
redundant work that it does compared to Blender (see T94644) and
probably prepare the ground for Vulkan support. As a matter of fact,
this patch also removes the OpenGL initialization that OpenSubdiv would
do here. This removal is not related to the bug fix, but necessary to not
have to copy more files/code over.
Differential Revision: https://developer.blender.org/D14131
std::min was used without including the algorithm
header. Seems to be implicitly included by
something in newer MSVC versions and GCC, however
vs16.4 needed a little help here.
This evaluator is used in order to evaluate subdivision at render time, allowing for
faster renders of meshes with a subdivision surface modifier placed at the last
position in the modifier list.
When evaluating the subsurf modifier, we detect whether we can delegate evaluation
to the draw code. If so, the subdivision is first evaluated on the GPU using our own
custom evaluator (only the coarse data needs to be initially sent to the GPU), then,
buffers for the final `MeshBufferCache` are filled on the GPU using a set of
compute shaders. However, some buffers are still filled on the CPU side, if doing so
on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose
logic is hardly GPU compatible).
This is done at the mesh buffer extraction level so that the result can be readily used
in the various OpenGL engines, without having to write custom geometry or tesselation
shaders.
We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in
order to control the data layout, and interpolation. For example, we store vertex colors
as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float
types.
In order to still access the modified geometry on the CPU side, for use in modifiers
or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`.
Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will
create such a wrapper if possible. If the final subdivision surface is not needed on
the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used.
Enabling or disabling GPU subdivision can be done through the user preferences (under
Viewport -> Subdivision).
See patch description for benchmarks.
Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport
Differential Revision: https://developer.blender.org/D12406
Using the `MEM_*` API from C++ code was a bit annoying:
* When converting C to C++ code, one often has to add a type cast on
returned `void *`. That leads to having the same type name three times
in the same line. This patch reduces the amount to two and removes the
`sizeof(...)` from the line.
* The existing alternative of using `OBJECT_GUARDED_NEW` looks a out
of place compared to other allocation methods. Sometimes
`MEM_CXX_CLASS_ALLOC_FUNCS` can be used when structs are defined
in C++ code. It doesn't look great but it's definitely better. The downside
is that it makes the name of the allocation less useful. That's because
the same name is used for all allocations of a type, independend of
where it is allocated.
This patch introduces three new functions: `MEM_new`, `MEM_cnew` and
`MEM_delete`. These cover the majority of use cases (array allocation is
not covered).
The `OBJECT_GUARDED_*` macros are removed because they are not
needed anymore.
Differential Revision: https://developer.blender.org/D13502
