multi_draw_indirect_counter #594
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This code is working now, but it wasn't working an hour ago, and I didn't really do anything to fix it. I want to make sure I understand what's going on before this is merged. But it's rather straightforward code. |
fyellin
commented
Sep 20, 2024
…d up inside a function definition.
|
The naga bug is back. |
… works when we do the bug_patch, but not otherwise.
Add missing newline
|
This code is waiting for a new wgpu-native-lib to include: gfx-rs/wgpu-native#432 We pull directly from wpgu-native, so their next update should include the change. I'm not sure how wgpu-native decides which version of wgpu to use. |
Here's an example: gfx-rs/wgpu-native#402 Basically bumping versions in |
|
I ask because it seems that tag 22.1.0.5 was created on gfx-rs/wgpu-native on Sept 12. Meanwhile I don't see any tags on gfx-rs/wgpu, and the fix was merged into "trunk" on Sept 2. But it doesn't seem to have made it into the build of lib-native. |
|
That's correct. The latest tag from wgpu-core is 22.1.0. The versioning of wgpu-native is to take the version of wgpu-core and add a fourth number to it. We've had a few releases (22.1.0.x) for bugfixes and changes to how the binary archives are composed. |
|
I've confirmed that this code works with a patched wgpu. I've tried it out both on emulated GPUs and on real GPUs. I just need to wait for us to catch up. |
…ime sleeping. (pygfx#655) * Bring sleep back to WgpuAwaitable. Slowly back off on the amount of time sleeping. * Need anyio * Need anyio * Need anyio * Remove no-longer-needed-comment
…ygfx#657) * Stop setting AA_EnableHighDpiScaling since Qt Says it is not needed See https://doc.qt.io/qt-6/highdpi.html * update comment * fix variable name * Update wgpu/gui/qt.py
The pytest anyio plugin is built into anyio. You don't need this package. https://github.com/graingert/pytest-anyio/blob/default/pytest_anyio.py
* Actually run the tests instead of fully skipping them Its easy to "fix tests" by completely skipping them ^_^_^_^_^ * Add the gui_auto.png * TODO: actually fix --- brutally force the cube example to run so i can regenereate the images * Attempt at better testing???
* Improve error reporting with screenshots and provide a method to test locally * improve readme. * Break everything because i don't understand async * run ruff format... * spelling * minimize changes
* use default lavapipe for screenshot * ruff format * bump pypy to 3.10 * run codegen
Maybe the cause of pygfx/pygfx#938 I'm still getting: ``` Traceback (most recent call last): File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/wgpu/gui/_gui_utils.py", line 76, in log_exception yield File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/wgpu/gui/base.py", line 167, in _draw_frame_and_present self.draw_frame() File "/home/mark/git/python-owl/owl/gui/_scene_control_gfx.py", line 2464, in animate self._renderer.render( File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/renderer.py", line 536, in render container_group = get_pipeline_container_group(wobject, environment) File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/pipeline.py", line 181, in get_pipeline_container_group pipeline_container_group.update(wobject, environment, changed_labels) File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/pipeline.py", line 261, in update container.update(wobject, environment, env_hash, changed) File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/pipeline.py", line 336, in update raise err File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/pipeline.py", line 333, in update self.update_shader_data(wobject, changed) File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/pipeline.py", line 369, in update_shader_data self.update_bind_groups() File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/pipeline.py", line 464, in update_bind_groups binding_des, binding_layout_des = binding.get_bind_group_descriptors( File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/binding.py", line 134, in get_bind_group_descriptors "resource": ensure_wgpu_object(resource), File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/pygfx/renderers/wgpu/engine/update.py", line 173, in ensure_wgpu_object resource._wgpu_object = wgpu_texture.create_view() File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/wgpu/backends/wgpu_native/_api.py", line 2343, in create_view id = libf.wgpuTextureCreateView(self._internal, struct) File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/wgpu/backends/wgpu_native/_helpers.py", line 384, in proxy_func error_type_msg = error_handler.release(name) File "/home/mark/miniforge3/envs/dev/lib/python3.10/site-packages/wgpu/backends/wgpu_native/_helpers.py", line 321, in release n = self._proxy_stack.pop() IndexError: pop from an empty deque ``` Co-authored-by: Almar Klein <[email protected]>
* Improve docs regarding drives on Linux * tweak more * Add note on wsl * not everyone uses pip
For the conda-forge packages I package wgpu-native myself and install them in the "system" libraries. This should continue to allow you to use the "wheel" vendored headers but then also look for the headers in the "system" locations. I've been dragging on a simpler (less upstreamable) version of this patch for about a year but I'm hoping I don't have to with this addition.
* Allow disabling downloading lib when creating wheel * Update tools/hatch_build.py Co-authored-by: Mark Harfouche <[email protected]> * lazy-import * tweak --------- Co-authored-by: Mark Harfouche <[email protected]>
* stats * stats * tweak
* stats auto render * auto render property
* add simple matmul example * flake8, put shader src in same file
* pass constants to compute helper * flake8 remove whitespace
* Improve performance for 'bitmap' render method * add comment * codegen * Only store buffer, and share between textures * fix memtest * codegen * leftover print call
* Release notes * --amend * more correct * adapt to new ruff
* update headers * try to fix hparser * handle escaped newlines * run codegen * fix escaped newlines in lib * add string view in logger * simplify limits * fix more callbacks * update surface descriptors * update shader module * update render pipeline * fix surface status * update copy info * fix parsing of MAX values * update buffer map * avoid magic values * update generate_report * fix limits again * fix labels * update error messages * make async unimplemented * fix string view null value * enable bundle test * avoid pointer references * request native limits * update mem tests * cleanup hparser * handle nested structs * update codegen test * apidiff compilation hint * lint * ruff format * bump wgpu-native build * set callback modes * small tweaks * tweak codegen of flags so they're not reported missing * Make limit logic a bit more consistent * tweak new to_c_string_view * tweak to_c_string_view and add from_c_string_view * Check result of function that return WGPUStatus * Need the null-string * map-read workaround still needed. Update test/comments a bit * comment on textureview.usage * codegen * the new test failed intermittendly * Use to_c_string_view for labels too * avoid hardcoded enum value --------- Co-authored-by: Almar Klein <[email protected]>
* new idl * tweaks to codegen * Codegen and adjustements * remove apidiff for create_shader_module * small tweaks * lint
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Implementation of the final two functions.
Fix some typos I had previous made.