This merge the description into one struct only that can be more easily
copied during `finalize()`.
The in and out layout parameters are better named and extended with the
invocation count (with fallback support)
This is a first part of the Shader Create Info system could be.
A shader create info provides a way to define shader structure, resources
and interfaces. This makes for a quick way to provide backend agnostic
binding informations while also making shader variations easy to declare.
- Clear source input (only one file). Cleans up the GPU api since we can create a
shader from one descriptor
- Resources and interfaces are generated by the backend (much simpler than parsing).
- Bindings are explicit from position in the array.
- GPUShaderInterface becomes a trivial translation of enums and string copy.
- No external dependency to third party lib.
- Cleaner code, less fragmentation of resources in several libs.
- Easy to modify / extend at runtime.
- no parser involve, very easy to code.
- Does not hold any data, can be static and kept on disc.
- Could hold precompiled bytecode for static shaders.
This also includes a new global dependency system.
GLSL shaders can include other sources by using #pragma BLENDER_REQUIRE(...).
This patch already migrated several builtin shaders. Other shaders should be migrated
one at a time, and could be done inside master.
There is a new compile directive `WITH_GPU_SHADER_BUILDER` this is an optional
directive for linting shaders to increase turn around time.
What is remaining:
- pyGPU API {T94975}
- Migration of other shaders. This could be a community effort.
Reviewed By: jbakker
Maniphest Tasks: T94975
Differential Revision: https://developer.blender.org/D13360
A recent security update to macOS 10.15.7 causes crashes when using Eevee and
various other 3D viewport features. It appears that glGenerateMipmap is
broken, causing a crash whenever its commands are flushed/submitted to the GPU.
Ideally this would be fixed in a driver update, however it's unlikely this will
happen. Earlier macOS versions have been receiving security updates for 2 years,
and that window has just passed for 10.15. Further, computers with these GPUs
can't upgrade to a newer macOS version.
As a workaround, disable mipmaps on these GPUs, by setting the mipmap max level
to 0 and not calling glGenerateMipmaps. Effects like depth of field also use
mipmaps, but fill in the mip levels by other means. In those cases we keep the
mipmap level.
Differential Revision: https://developer.blender.org/D13295
This wraps the functionality used to speedup EEVEE volumetrics.
This touches the rendering code of EEVEE as it should fix a mis-usage of
the GL barrier. The barrier changed type and location, removing an
unused barrier.
This makes it possible to disable all the extensions when forcing
workarounds.
Also it will allow future options to selectively disable each extension
to know which one is buggy.
This is to improve debugging on older hardware that may not support
4.3 debug capabilities (like Macs).
This avoids sprinkling glGetErrors manually. This might still be needed
to find the root cause since not all functions are covered.
This overrides the functions pointers that GLEW have already init.
This is only enabled if using --debug-gpu option and the debug extension
are not available.
This also cleanup the usage of GLContext::debug_layer_support and use
wrapper to set object labels.
This makes the GPUContext follow the same naming convention as the rest
of the module.
Also add a static getter for extra bonus style (no need for casts):
- Context::get()
- GLContext::get()
This is part of the Vulkan task T68990.
This introduce a new GLQueryPool for managing queries in an
implementation agnostic manner.
This modify the GPU selection query to use this new object.
This also make use of blender::Vector for better code quality.
No real functionnal change.
This is part of the Vulkan task T68990.
This commits changes a few things:
- Rename extensions to capabilities (but left the file name untouched).
- Cubemap mip render workaround detection is rewritten using gl
commands to avoid using the GPU API before initialization.
- Put all the capabilities that are only relevant for the GL backend
inside GLContext as static variables.
- Cleanup the names of the limit variables.
- Separate all GL related workaround search inside the GL module.
This is a massive cleanup needed for vulkan support T68990. It provides:
- More meaningful enums with conversion functions.
- Less hacky supports of arrays and cubemaps (all considered layered).
- More inline with the stateless design of vulkan and modern GL.
- Methods Fallbacks are using framebuffer functions that are wrapped
instead of implementing inside the texture module.
What is not in there:
- API change.
- Samplers support (breaks a few effects).
# Conflicts:
# source/blender/gpu/GPU_texture.h
This is in preparation of vulkan backend. We move all opengl
functionnalities behind an abstract class.
This also cleansup the "dynamic" ubo create and rename it to
`GPU_uniformbuf_from_list()`
Contains, no functional change.
Part of T68990 Vulkan support.
This changes the drawing paradigm a bit. The VAO configuration is done
JIT-style and depends on context active shader.
This is to allow more flexibility for implementations to do optimization
at lower level.
The vao cache is now its own class to isolate the concept. It is this
class that is reference by the GLContext for ownership of the containing
VAO ids.
This just set a global object responsible for allocating new objects
in a thread safe way without needing any GPUContext bound to this
thread.
This also introduce the GLContext which will contain all the GL related
functions for the current context.
# Conflicts:
# source/blender/gpu/intern/gpu_context.cc
This just set a global object responsible for allocating new objects
in a thread safe way without needing any GPUContext bound to this
thread.
This also introduce the GLContext which will contain all the GL related
functions for the current context.
