This PR moves the responsibility of destroying discarded resources to
the submission thread. Previous implementation could be blocked and
would not always run.
This solves memory leak when rendering in background and keeps the
overall memory usage lower as all is done in a single location.
Pull Request: https://projects.blender.org/blender/blender/pulls/144440
With 5.0 we start requiring this extension for GL and VK.
All of our target hardware supports it with up to date
drivers.
Some old drivers were disabling this extension because of
buggy behavior. We simply drop support for them in 5.0.
This allows us to remove a lot of code and the last
shader create info override done at startup. This will
unlock more refactoring of the shader create info into
static classes to reduce binary size and other benefits.
## TODO:
- [x] Remove checks for ARB_shader_draw_parameters
- [x] Remove checks for ARB_clip_control
- [x] Check for the extension on startup for OpenGL
- [x] Check for the extension on startup for Vulkan
- [x] ~~Add user facing popup message about minimum
requirements not being met.~~ Done using the same
popup as old hardware.
Pull Request: https://projects.blender.org/blender/blender/pulls/142334
This PR will swap device local memory to system ram. It relies on
VK_EXT_external_memory and VK_EXT_pageable_device_local_memory
extensions to be supported by the system.
Most platforms support these extensions.
Pull Request: https://projects.blender.org/blender/blender/pulls/144422
There are stability issues with the vulkan backend. Some scenes indicate
that there is no space left to allocate the next descriptor buffer. For
stability reasons we will disable descriptor buffers and look into a
better solution.
Pull Request: https://projects.blender.org/blender/blender/pulls/144421
Use signed range and indices to avoid dereferencing tokens
before the start of the Token array.
# Conflicts:
# source/blender/gpu/glsl_preprocess/shader_parser.hh
This removes one limitation of the guarded scopes that needed to
not have return statement. Now a dummy return statement is created.
See #139233 for original implementation.
This adds the following features:
- `class` keyword support: checked by C++, mutated to struct for shader.
- `private` and `public` keywords: checked by C++, removed for shader.
- `static` methods.
- `const` and non-const methods.
What is not supported:
- Constructors
- Destructors
- operators
- Method definition outside of class definition
- member reference without `this` keyword.
This is implemented using a very simple lexer/parser allowing semantic traversal.
Pull Request: https://projects.blender.org/blender/blender/pulls/144025
Swapchain handling of minimized windows wasn't correct. On some
platforms it still tried to create images with no surface.
This PR will discard swapchains of minimized windows, but still being
able to flush the render graph and free resources.
Pull Request: https://projects.blender.org/blender/blender/pulls/144189
Vulkan 1.0 render passes have been replaced by dynamic rendering in 1.2.
Blender Vulkan backend was implemented with dynamic rendering in mind.
All our supported platforms support dynamic rendering. Render pass support
was added to try to work around an issue with legacy drivers. However these
drivers also fail with render passes.
Using render passes had several limitations (blending and some workbench
features were not supported). As no GPU uses it and it is quite some code
to support it is better to remove it.
Pull Request: https://projects.blender.org/blender/blender/pulls/144149
The issue in #142046 is caused by the deferred lighting pass not
running.
Despite plenty of testing and debugging we have not been able to spot
any errors in the Blender code.
Checking the link status of specialization shader variants seems to fix
the issue, which may suggest a bug on the driver side.
Pull Request: https://projects.blender.org/blender/blender/pulls/144150
On some drivers, the GLSL compiler doesn't reflect the omitted
`local_size_*` of a compute shader inside `gl_WorkGroupSize`.
This lead to the 2D size computation of 1D workgroups to become
0 which was bypassing the parallel reduction algorithms.
Ensuring `local_size_*` are always set fixes the issue.
For clarity, also fix the 1D shaders to not use `gl_WorkGroupSize.y`.
This also fix a copy paste error in the Metal backend.
This issue affected AMD drivers on Windows.
Rel #142046
Candidate for backporting to 4.5 LTS.
Pull Request: https://projects.blender.org/blender/blender/pulls/144056
Resetting and recycling of immediate drawing buffers was never done and
would leak memory as the buffers were only destroyed when Blender
exited.
This is solved by not recycling or resetting the buffers and rely on the
discard pools. Additional cleanup of removing unused code-paths is also
part of this change so it can be backported to 4.5.
Pull Request: https://projects.blender.org/blender/blender/pulls/143995
The Map UV node does not work when the UV input is a single value, where
it is expected that the output will also be single value. This was
simply not implemented for GPU, so this patch does that.
Pull Request: https://projects.blender.org/blender/blender/pulls/143096
This assert was to notify developers that they are using a framebuffer
configuration that isn't supported when using render passes. Render
passes should not be used and will be removed later in 5.0. Removing the
assert will already help during triaging.
Pull Request: https://projects.blender.org/blender/blender/pulls/143936
`delete_cached_binary` didn't work correctly since the cache hash could
be overwritten before the subprocess crashed.
This fix requires an extra memory copy, but it doesn't seem to incur any
measurable overhead.
An alternative could be to store the cache name in its own variable in
ShaderBinaryHeader, but that would break cache files compatibility.
Pull Request: https://projects.blender.org/blender/blender/pulls/143649
When creating `GPUSecondaryContext`s, `epoxy_gl_version` returns 0 and
`epoxy_has_gl_extension` always returns false.
This is caused by `GPU_context_create` being called without the ghost
context being activated.
Activating it fixes the issue.
Pull Request: https://projects.blender.org/blender/blender/pulls/142715
Shader compilation no longer uses the `WM_job` API.
Add a `GPU_shader_batch_is_compiling` function to query if there's any
shader compilation happening, and update `bpy_app_is_job_running` to
handle this as a special case.
Pull Request: https://projects.blender.org/blender/blender/pulls/143559
It's broken and it's not needed anymore, since we require OpenGL 4.3.
Fix#142620
Note: Debug callbacks still won't work for compilation threads,
but that's relatively harmless and should cause no issues for the user.
This will be fixed in 5.0 with #142715.
(4.5 backport candidate)
Pull Request: https://projects.blender.org/blender/blender/pulls/142716
Recursive downsample was only used by workbench DoF
which can be expressed using mip view instead.
The mip render workaround was creating GL error on startup
and is not needed anymore.
Pull Request: https://projects.blender.org/blender/blender/pulls/143246
It allows to implement tricks based on a knowledge whether the path
ever cam through a portal or not, and even something more advanced
based on the number of portals.
The main current objective is for strokes shading: stroke shader
uses Ray Portal BSDF to place ray to the center of the stroke and
point it in the direction of the surface it is generated for. This
gives stroke a single color which matches shading of the original
object. For this usecase to work the ray bounced from the original
surface should ignore the strokes, which is now possible by using
Portal Depth input and mixing with the Transparent BSDF. It also
helps to make shading look better when there are multiple stroke
layers.
A solution of using portal depth is chosen over a single flag due
to various factors:
- Last time we've looked into it it was a bit tricky to implement
as a flag due to us running out of bits.
- It feels to be more flexible solution, even though it is a bit
hard to come up with 100% compelling setup for it.
- It needs to be slightly different from the current "Is Foo"
flags, and be more "Is Portal Descendant" or something.
An extra uint16 is added to the state to count the portal depth,
but it is only allocated for scenes that use Ray Portal BSDF.
Portal BSDF still increments Transparent bounce, as it is required
to have some "limiting" factor so that ray does not get infinitely
move to different place of the scene.
Ref #125213
Pull Request: https://projects.blender.org/blender/blender/pulls/143107
When using GPU_DEVICE_QUALCOMM video drivers on Win32 with OpenGL (just
ARM) if the driver is too old a console message says that the driver is
not capable of running "Blender 4.0". This PR just adds "or later".
Pull Request: https://projects.blender.org/blender/blender/pulls/143108
This includes a new list structure type and socket shape, a node
to create lists, a node to retrieve values from lists, and a node to
retrieve the length of lists. It also implements multi-function support
so that function nodes work on lists.
There are three nodes included in this PR.
- **List** Creates a list of elements with a given size. The values
are computed with a field that can use the index as an input.
- **Get List Item** A field node that retrieves an element from a
a list at a given index. The index input is dynamic, so if the input
is a list, the output will be a list too.
- **List Length** Just gives the length of a list.
When a function node is used with multiple list inputs, the shorter
lists are repeated to extend it to the length of the longest.
The list nodes and structure type are hidden behind an experimental
feature until we can be sure they're useful for an actual use case.
Pull Request: https://projects.blender.org/blender/blender/pulls/140679
This commit renames the GHOST Metal graphic context class and related
files / references from `GHOST_ContextCGL` to `GHOST_ContextMTL`. When
the Metal backend was first introduced, this context file contained
both the old OpenGL context (CGL, for macOS Core OpenGL API), and the
newer Metal context. Since #110185 all CGL related code was removed
from the class, making it Metal only, and thus rendering the old class
name outdated and potentially misleading. In addition to the rename,
unused OpenGL related forward declarations and old TODOs were also
removed.
Pull Request: https://projects.blender.org/blender/blender/pulls/142519
This allows to store a number of vertex to draw
per batch without specifying any attribute.
This allows to create batches that are empty but
still holds the amount of geometry to produce.
Needed for new curve drawing #142969.
Pull Request: https://projects.blender.org/blender/blender/pulls/143052
This was caused by the hardware derivatives output being affected
by render resolution. Scaling them back to the full resolution
value fixes the issue.
This also fixes the Wireframe node that also relies on derivatives.
Pull Request: https://projects.blender.org/blender/blender/pulls/142101
