Commit Graph
1962 Commits
Author SHA1 Message Date
teoxoy 0e1c1f7c07 replace the tracker in Device with a new DeviceTracker 2024-07-09 10:48:52 +02:00
teoxoy 3142e15907 remove the triage suspected machinery 2024-07-09 10:48:52 +02:00
teoxoy 3fba403060 don't check if the buffer is still "present" from the user's perspective in LifetimeTracker.handle_mapping
This change doesn't change behavior as `Global.buffer_drop` already unmaps the buffer.
2024-07-09 10:48:52 +02:00
teoxoy 3cc6c2743a remove LifetimeTracker.future_suspected_{buffers,textures} 2024-07-09 10:48:52 +02:00
teoxoy 61739d9583 remove PendingWrites.executing_command_buffers
The Vec only ever contained 0 or 1 command buffers.
We now acquire an encoder on every submit for pending writes but that's not a problem since those are coming from a pool anyway.
2024-07-09 10:48:52 +02:00
teoxoy 439e28bfc1 move trackers into EncoderInFlight on submit 2024-07-09 10:48:52 +02:00
teoxoy 140495006e remove device_maintain_ids 2024-07-09 10:48:52 +02:00
teoxoy c6761bdd7a add ActiveSubmission.temp_resources that contains all temporary resources.
It's worth noting that `suspected_resources` never contained those resources.
2024-07-09 10:48:52 +02:00
teoxoy 7223bfa88d remove Arcs around TempResource variants 2024-07-09 10:48:52 +02:00
teoxoy ef909d0a82 remove Device.create_buffer_impl 2024-07-09 10:48:52 +02:00
teoxoy ab88dce759 remove unused TempResource::Buffer 2024-07-09 10:48:52 +02:00
teoxoy fabbca294a use StagingBuffer instead of Buffer for mapped_at_creation Buffers 2024-07-09 10:48:52 +02:00
teoxoy 5266bd1f08 change prepare_staging_buffer to return a non null u8 pointer 2024-07-09 10:48:52 +02:00
teoxoy bd2b284a26 remove unused TempResource::Texture 2024-07-09 10:48:52 +02:00
Nicolas SilvaandAndreas Reich 4c6318c0d2 Expose gpu allocation configuration options (#5875)
* Expose gpu allocation configuration options

This commit adds hints to control memory allocations strategies to the configuration options. These hints allow for automatic profiles such as optimizing for performance (the default, makes sense for a game), optimizing for memory usage (typically more useful for a web browser or UI library) and specifying settings manually.

The details of gpu allocation are still in flux. The goal is to switch vulkan and metal to gpu_allocator which is currently used with d3d12. gpu_allocator will also likely receive more configuration options, in particular the ability to start with smaller memory block sizes and progressively grow the block size. So the manual settings already provision for this upcoming option. Another approach could be to wait and add the manual option after the dust settles.

The reason for providing presets and defining values in the backends is that I am convinced that optimal fonigurations should take hardware capabilities into consideration. It's a deep rabbithole, though, so that will be an exercise for later.

* changelog

* Update CHANGELOG.md

Co-authored-by: Andreas Reich <r_andreas2@web.de>

* Add a comment about not entirely knowing what we are doing

---------

Co-authored-by: Andreas Reich <r_andreas2@web.de>
2024-07-08 14:49:44 +02:00
Erich Gubler d41b9ab29f refactor: use built-in {integer}::power_of_two methods (#5909) 2024-07-04 23:34:19 -04:00
teoxoy 7910fd8059 [wgpu-hal] require a Surface to be passed to Instance.enumerate_adapters() on WebGL2
Also makes wgpu's `enumerate_adapters()` native only.
2024-07-04 10:17:07 +02:00
Imbris 3a6814770a Allow unconsumed inputs in fragment shaders (#5531)
* Allow unconsumed inputs in fragment shaders by removing them from vertex
outputs when generating HLSL.

Fixes https://github.com/gfx-rs/wgpu/issues/3748

* Add naga::back::hlsl::FragmentEntryPoint for providing information
  about the fragment entry point when generating vertex entry points via
  naga::back::hlsl::Writer::write. Vertex outputs not consumed by the
  fragment entry point are omitted in the final output struct.
* Add naga snapshot test for this new feature,
* Remove Features::SHADER_UNUSED_VERTEX_OUTPUT,
  StageError::InputNotConsumed, and associated validation logic.
* Make wgpu dx12 backend pass fragment shader info when generating
  vertex HLSL.
* Add wgpu regression test for allowing unconsumed inputs.

* Address review

* Add note that nesting structs for the inter-stage interface can't
  happen.
* Remove new TODO notes (some addressed and some transferred to an issue
  https://github.com/gfx-rs/wgpu/issues/5577)
* Changed issue that regression test refers to 3748 -> 5553
* Add debug_assert that binding.is_some() in hlsl writer
* Fix typos caught in CI

Also, fix compiling snapshot test when hlsl-out feature is not enabled.
2024-07-04 09:08:46 +02:00
teoxoy e26d2d7763 move command buffer resolving in Global's methods 2024-07-03 15:01:41 +02:00
teoxoy a9c74f42d6 use Surface.get_capabilities in all relevant places 2024-07-03 15:01:41 +02:00
teoxoy ef7d27233e use a read guard instead of write 2024-07-03 15:01:41 +02:00
teoxoy eb0eb342cf unregister adapters like all other resources 2024-07-03 15:01:41 +02:00
teoxoy 5b2e78da45 cleanup adapter usages within the Device 2024-07-03 15:01:41 +02:00
teoxoy 4f6223534f take guard to render bundles at the top of resolve_render_command_ids 2024-07-03 15:01:41 +02:00
teoxoy addd80ccd3 replace Registry.read().get_owned() with Registry.get() 2024-07-03 15:01:41 +02:00
teoxoy 3df4c8a502 remove unused std::ops::Index impl for Storage 2024-07-03 15:01:41 +02:00
teoxoy 9308e15048 use Storage.get_owned in cases where we used to later clone the resource anyway 2024-07-03 15:01:41 +02:00
teoxoy 249c8023ff simplify State.set_bind_group 2024-07-03 15:01:41 +02:00
teoxoy dbcb273188 remove StatelessTracker.add_single 2024-07-03 15:01:41 +02:00
teoxoy 39a268c624 remove FutureId.assign_existing 2024-07-03 15:01:41 +02:00
teoxoy cd9f003477 move Device.set_queue call in create_device_and_queue_from_hal 2024-07-03 15:01:41 +02:00
teoxoy 96255d484f change FutureId.assign to not return a clone of the resource passed to it 2024-07-03 15:01:41 +02:00
teoxoy 7f3e1bba14 move code after calls to FutureId.assign inside Device fns 2024-07-03 15:01:41 +02:00
teoxoy 40022c1584 remove duplicate check, it's already present in Device.create_texture_view 2024-07-03 15:01:41 +02:00
teoxoy dc55cb436c add Texture::new fn 2024-07-03 15:01:41 +02:00
teoxoy 53c79bdf7a move buffer creation logic in a new Device method 2024-07-03 15:01:41 +02:00
teoxoy 029168f14a remove no-op, calling schedule_resource_destruction with last_submit_index: !0 does nothing 2024-07-03 15:01:41 +02:00
teoxoy a971e7f7d4 remove duplicate check, it's already present in Device.create_buffer 2024-07-03 15:01:41 +02:00
teoxoy 7b387007a8 make Device.create_buffer return an Arc<Buffer> 2024-07-03 15:01:41 +02:00
teoxoy b8c5b2275e don't clone resources for ResourceMetadataProvider::Direct 2024-07-03 15:01:41 +02:00
teoxoy 69b44c6a32 gate fns behind either serde or replay since Firefox needs those 2024-07-02 18:06:19 +02:00
teoxoy b61be30e53 move out ID to resource mapping code from Device.create_bind_group 2024-07-02 18:06:19 +02:00
teoxoy 1be51946e3 move out ID to resource mapping code from pipeline creation methods 2024-07-02 18:06:19 +02:00
teoxoy d8b0b5975d assign resources/errors in the Global's pipeline creation functions 2024-07-02 18:06:19 +02:00
teoxoy 79b15b8de8 move out ID to BGL mapping from Device.create_pipeline_layout() 2024-07-02 18:06:19 +02:00
teoxoy 1f6ac35e83 change Storage.force_replace to take an Arc<T> 2024-07-02 18:06:19 +02:00
teoxoy 896418c740 Panic on non-existent entry in Storage.get, this is consistent with Element::Vacant 2024-07-02 18:06:19 +02:00
teoxoy 16c7d84d63 remove Registry.try_get() 2024-07-02 18:06:19 +02:00
teoxoy b904e2c759 use Registry.get() for device_set_device_lost_closure 2024-07-02 18:06:19 +02:00
teoxoy c4d71d9af5 use Registry.get() instead of Registry.read().get_owned() 2024-07-02 18:06:19 +02:00