wgpu: Some optimisations, some refactors, much less buffers#14919

Merged

Dinnerbone merged 18 commits intoruffle-rs:masterfrom Dinnerbone:staging_belt

Jan 27, 2024

+1,264-1,533

Commits on Jan 27, 2024

wgpu: Add limits to BufferBuilder (but currently not respected for Meshes)
Dinnerbone
committed
wgpu: Try using a single buffer/bindgroup for transform/colortransforms, no longer dynamically allocated
Dinnerbone
committed
wgpu: Remove push constant support - simplifies code significantly, and it should be same perf now
Dinnerbone
committed
wgpu: Use max_uniform_buffer_binding_size from adapter
Dinnerbone
committed
wgpu: Extract out ESTIMATED_OBJECTS_PER_CHUNK, set it to 200
Dinnerbone
committed
wgpu: Merge Transforms and ColorAdjustments
Dinnerbone
committed
wgpu: Remove uniform encoder
Dinnerbone
committed
wgpu: Make sync index Option<> in make_queue_sync_handle
Dinnerbone
committed
wgpu: Use staging belt and single buffer for filter arguments
Dinnerbone
committed
wgpu: Use vertex buffer and staging belt for simple filters
Dinnerbone
committed
wgpu: Use vertex buffer and staging belt for glow filter
Dinnerbone
committed
wgpu: Use vertex buffer and staging belt for bevel filter
Dinnerbone
committed
wgpu: Clean up DynamicTransforms, after the merge
Dinnerbone
committed
wgpu: Remove ouroboros and typed-arena, no longer needed
Dinnerbone
committed
wgpu: Don't use max_uniform_buffer_binding_size as a limit for BufferBuilder when making meshes
Dinnerbone
committed
swf: Derive Hash for Gradient and associated types
Dinnerbone
committed
render: Make Mesh a struct, not just an alias for Vec<Draw>
Dinnerbone
committed
wgpu: Deduplicate common gradients from a single Mesh
Dinnerbone
committed