Erroneous cache warming case where the generated material is
identical to default material and cached shader is re-used,
resulting in case where the parent shader is identical to the
source.
Authored by Apple: Michael Parkin-White
Ref #96261
This patch adds initial support for compute shaders to
the vulkan backend. As the development is oriented to the test-
cases we have the implementation is limited to what is used there.
It has been validated that with this patch that the following test
cases are running as expected
- `GPUVulkanTest.gpu_shader_compute_vbo`
- `GPUVulkanTest.gpu_shader_compute_ibo`
- `GPUVulkanTest.gpu_shader_compute_ssbo`
- `GPUVulkanTest.gpu_storage_buffer_create_update_read`
- `GPUVulkanTest.gpu_shader_compute_2d`
This patch includes:
- Allocating VkBuffer on device.
- Uploading data from CPU to VkBuffer.
- Binding VkBuffer as SSBO to a compute shader.
- Execute compute shader and altering VkBuffer.
- Download the VkBuffer to CPU ram.
- Validate that it worked.
- Use device only vertex buffer as SSBO
- Use device only index buffer as SSBO
- Use device only image buffers
GHOST API has been changed as the original design was created before
we even had support for compute shaders in blender. The function
`GHOST_getVulkanBackbuffer` has been separated to retrieve the command
buffer without a backbuffer (`GHOST_getVulkanCommandBuffer`). In order
to do correct command buffer processing we needed access to the queue
owned by GHOST. This is returned as part of the `GHOST_getVulkanHandles`
function.
Open topics (not considered part of this patch)
- Memory barriers & command buffer encoding
- Indirect compute dispatching
- Rest of the test cases
- Data conversions when requested data format is different than on device.
- GPUVulkanTest.gpu_shader_compute_1d is supported on AMD devices.
NVIDIA doesn't seem to support 1d textures.
Pull-request: #104518
This reverts commit 19222627c6.
Something went wrong here, seems like this commit merged the main branch
into the release branch, which should never be done.
Certain material node graphs can be very expensive to run. This feature aims to produce secondary GPUPass shaders within a GPUMaterial which provide optimal runtime performance. Such optimizations include baking constant data into the shader source directly, allowing the compiler to propogate constants and perform aggressive optimization upfront.
As optimizations can result in reduction of shader editor and animation interactivity, optimized pass generation and compilation is deferred until all outstanding compilations have completed. Optimization is also delayed util a material has remained unmodified for a set period of time, to reduce excessive compilation. The original variant of the material shader is kept to maintain interactivity.
Also adding a new concept to gpu::Shader allowing assignment of a parent shader from which a shader can pull PSO descriptors and any required metadata for asynchronous shader cache warming. This enables fully asynchronous shader optimization, without runtime hitching, while also reducing runtime hitching for standard materials, by using PSO descriptors from default materials, ahead of rendering.
Further shader graph optimizations are likely also possible with this architecture. Certain scenes, such as Wanderer benefit significantly. Viewport performance for this scene is 2-3x faster on Apple-silicon based GPUs.
Authored by Apple: Michael Parkin-White
Ref T96261
Pull Request #104536
This replaces `GPU_SHADER_3D_POINT_FIXED_SIZE_VARYING_COLOR` by
GPU_SHADER_2D_POINT_UNIFORM_SIZE_UNIFORM_COLOR_OUTLINE_AA`.
None of the usage made sense to not use the AA shader.
Scale the point size to account for the rounded shape.
The GPU module has 2 different styles when reading back data from
GPU buffers. The SSBOs used a memcpy to copy the data to a
pre-allocated buffer. IndexBuf/VertBuf gave back a driver/platform
controlled pointer to the memory.
Readback is done for test cases returning mapped pointers is not safe.
For this reason we settled on using the same approach as the SSBO.
Copy the data to a caller pre-allocated buffer.
Reason why this API is currently changed is that the Vulkan API is more
strict on mapping/unmapping buffers that can lead to potential issues
down the road.
Pull Request #104571
Keep using the 3 evaluations dF_branch method for the Displacement output.
The optimized 2 evaluations method used by node_bump is now on its own macro (dF_branch_incomplete).
displacement_bump modifies the normal that nodetree_exec uses, so even with a refactor it wouldn’t be possible to re-use the computation anyway.
Avoid computing the non-derivative height twice.
The height is now computed as part of the main function, while the height at x and y offsets are still computed on a separate function.
The differentials are now computed directly at node_bump.
Co-authored-by: Miguel Pozo <pragma37@gmail.com>
Pull Request #104595
Documented all functions, adding use case and side effects.
Also replace the use of shortened argument name by more meaningful ones.
Renamed `GPU_batch_instbuf_add_ex` and `GPU_batch_vertbuf_add_ex` to remove
the `ex` suffix as they are the main version used (removed the few usage
of the other version).
Renamed `GPU_batch_draw_instanced` to `GPU_batch_draw_instance_range` and
make it consistent with `GPU_batch_draw_range`.
Blender was reporting that the GPU_TEXTURE_USAGE_HOST_READ wasn't set.
This is used to indicate that the textures needs to be read back to
CPU. Textures that don't need to be read back can be optimized by the
GPU backend.
Found during investigation of #104282.
`ED_GIZMO_CAGE2D_STYLE_CIRCLE` now draw circles. The previous `ED_GIZMO_CAGE2D_STYLE_CIRCLE`, which drew rectangles, is renamed to `ED_GIZMO_CAGE2D_STYLE_RECTANGLE`. The meaning of `ED_GIZMO_CAGE2D_STYLE_BOX` is now unclear and probably needs to be renamed too.
Ref T104280
Maniphest Tasks: T104280
Differential Revision: https://developer.blender.org/D17174
This patch adds support for compilation and execution of GLSL compute shaders. This, along with a few systematic changes and fixes, enable realtime compositor functionality with the Metal backend on macOS. A number of GLSL source modifications have been made to add the required level of type explicitness, allowing all compilations to succeed.
GLSL Compute shader compilation follows a similar path to Vertex/Fragment translation, with added support for shader atomics, shared memory blocks and barriers.
Texture flags have also been updated to ensure correct read/write specification for textures used within the compositor pipeline. GPU command submission changes have also been made in the high level path, when Metal is used, to address command buffer time-outs caused by certain expensive compute shaders.
Authored by Apple: Michael Parkin-White
Ref T96261
Ref T99210
Reviewed By: fclem
Maniphest Tasks: T99210, T96261
Differential Revision: https://developer.blender.org/D16990
Rewrite of the Workbench engine using C++ and the new Draw Manager API.
The new engine can be enabled in Blender `Preferences > Experimental > Workbench Next`.
After that, the engine can be selected in `Properties > Scene > Render Engine`.
When `Workbench Next` is the active engine, it also handles the `Solid` viewport mode rendering.
The rewrite aims to be functionally equivalent to the current Workbench engine, but it also includes some small fixes/tweaks:
- `In Front` rendered objects now work correctly with DoF and Shadows.
- The `Sampling > Viewport` setting is actually used when the viewport is in `Render Mode`.
- In `Texture` mode, textured materials also use the material properties. (Previously, only non textured materials would)
To do:
- Sculpt PBVH.
- Volume rendering.
- Hair rendering.
- Use the "no_geom" shader versions for shadow rendering.
- Decide the final API for custom visibility culling (Needed for shadows).
- Profile/optimize.
Known Issues:
- Matcaps are not loaded until they’re shown elsewhere. (e.g. when opening the `Viewort Shading` UI)
- Outlines are drawn between different materials of the same object. (Each material submesh has its own object handle)
Reviewed By: fclem
Maniphest Tasks: T101619
Differential Revision: https://developer.blender.org/D16826
Currently you can retrieve a mutable array from a const CustomData.
That makes code unsafe since the compiler can't check for correctness
itself. Fix that by introducing a separate function to retrieve mutable
arrays from CustomData. The new functions have the `_for_write`
suffix that make the code's intention clearer.
Because it makes retrieving write access an explicit step, this change
also makes proper copy-on-write possible for attributes.
Notes:
- The previous "duplicate referenced layer" functions are redundant
with retrieving layers with write access
- The custom data functions that give a specific index only have
`for_write` to simplify the API
Differential Revision: https://developer.blender.org/D14140
Currently the `MLoopUV` struct stores UV coordinates and flags related
to editing UV maps in the UV editor. This patch changes the coordinates
to use the generic 2D vector type, and moves the flags into three
separate boolean attributes. This follows the design in T95965, with
the ultimate intention of simplifying code and improving performance.
Importantly, the change allows exporters and renderers to use UVs
"touched" by geometry nodes, which only creates generic attributes.
It also allows geometry nodes to create "proper" UV maps from scratch,
though only with the Store Named Attribute node for now.
The new design considers any 2D vector attribute on the corner domain
to be a UV map. In the future, they might be distinguished from regular
2D vectors with attribute metadata, which may be helpful because they
are often interpolated differently.
Most of the code changes deal with passing around UV BMesh custom data
offsets and tracking the boolean "sublayers". The boolean layers are
use the following prefixes for attribute names: vert selection: `.vs.`,
edge selection: `.es.`, pinning: `.pn.`. Currently these are short to
avoid using up the maximum length of attribute names. To accommodate
for these 4 extra characters, the name length limit is enlarged to 68
bytes, while the maximum user settable name length is still 64 bytes.
Unfortunately Python/RNA API access to the UV flag data becomes slower.
Accessing the boolean layers directly is be better for performance in
general.
Like the other mesh SoA refactors, backward and forward compatibility
aren't affected, and won't be changed until 4.0. We pay for that by
making mesh reading and writing more expensive with conversions.
Resolves T85962
Differential Revision: https://developer.blender.org/D14365
**Changes**
As described in T93602, this patch removes all use of the `MVert`
struct, replacing it with a generic named attribute with the name
`"position"`, consistent with other geometry types.
Variable names have been changed from `verts` to `positions`, to align
with the attribute name and the more generic design (positions are not
vertices, they are just an attribute stored on the point domain).
This change is made possible by previous commits that moved all other
data out of `MVert` to runtime data or other generic attributes. What
remains is mostly a simple type change. Though, the type still shows up
859 times, so the patch is quite large.
One compromise is that now `CD_MASK_BAREMESH` now contains
`CD_PROP_FLOAT3`. With the general move towards generic attributes
over custom data types, we are removing use of these type masks anyway.
**Benefits**
The most obvious benefit is reduced memory usage and the benefits
that brings in memory-bound situations. `float3` is only 3 bytes, in
comparison to `MVert` which was 4. When there are millions of vertices
this starts to matter more.
The other benefits come from using a more generic type. Instead of
writing algorithms specifically for `MVert`, code can just use arrays
of vectors. This will allow eliminating many temporary arrays or
wrappers used to extract positions.
Many possible improvements aren't implemented in this patch, though
I did switch simplify or remove the process of creating temporary
position arrays in a few places.
The design clarity that "positions are just another attribute" brings
allows removing explicit copying of vertices in some procedural
operations-- they are just processed like most other attributes.
**Performance**
This touches so many areas that it's hard to benchmark exhaustively,
but I observed some areas as examples.
* The mesh line node with 4 million count was 1.5x (8ms to 12ms) faster.
* The Spring splash screen went from ~4.3 to ~4.5 fps.
* The subdivision surface modifier/node was slightly faster
RNA access through Python may be slightly slower, since now we need
a name lookup instead of just a custom data type lookup for each index.
**Future Improvements**
* Remove uses of "vert_coords" functions:
* `BKE_mesh_vert_coords_alloc`
* `BKE_mesh_vert_coords_get`
* `BKE_mesh_vert_coords_apply{_with_mat4}`
* Remove more hidden copying of positions
* General simplification now possible in many areas
* Convert more code to C++ to use `float3` instead of `float[3]`
* Currently `reinterpret_cast` is used for those C-API functions
Differential Revision: https://developer.blender.org/D15982
Required texture bytesize calculation for compacted data types was incorrectly calculated, resulting in an erroneous format conversion taking place instead of direct data upload.
Metal dummy buffer size also temporarily increased to address problematic cases where the bound buffer was too small for missing UBOs.
Authored by Apple: Michael Parkin-White
Ref T96261
Reviewed By: fclem
Maniphest Tasks: T96261
Differential Revision: https://developer.blender.org/D16904
(MacOS) only: In the System tab of the user preferences the user has the
ability to select a GPU backend that Blender will use. After changing
the GPU backend setting, the user has to restart Blender before the
setting is used.
It was added to start collecting feedback on the Metal backend without
using the command lines.
By default Blender will select OpenGL as backend. When Metal is selected
(via `--gpu-backend metal` or via user preferences) OpenGL will be used as
fallback when the platform isn't capable of running Metal.
- Support for non-contiguous shader resource bindings for all cases required by create-info
- Implement missing geometry shader alternative path for edit curve handle.
- Add support for non-float dummy textures to address all cases where default bindings may be required.
Authored by Apple: Michael Parkin-White
Ref T96261
Depends on D16721
Reviewed By: fclem
Differential Revision: https://developer.blender.org/D16777
Attributes are unifying around a name-based API, and we would like to
be able to move away from CustomData in the future. This patch moves
the identification of active and fallback (render) color attributes
to strings on the mesh from flags on CustomDataLayer. This also
removes some ugliness used to retrieve these attributes and maintain
the active status.
The design is described more here: T98366
The patch keeps forward compatibility working until 4.0 with
the same method as the mesh struct of array refactors (T95965).
The strings are allowed to not correspond to an attribute, to allow
setting the active/default attribute independently of actually filling
its data. When applying a modifier, if the strings don't match an
attribute, they will be removed.
The realize instances / join node and join operator take the names from
the first / active input mesh. While other heuristics may be helpful
(and could be a future improvement), just using the first is simple
and predictable.
Differential Revision: https://developer.blender.org/D15169
