[pull] main from gpuweb:main #1
Open
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
![pull[bot]](https://avatars.githubusercontent.com/in/12910?s=60&v=4){kind=small}
The section for an attribute has the following structure:
<grammar>
<valid values><rules for the parameters>
<description> <must-rules>
<describe what it is>
That puts the high level information last.
Reorder the sections so they have this structure:
<grammar>
<description> <describe what it is>
<must-rules>
<parameters><rules for the parameters>
I think this reads better
Co-authored-by: Mehmet Oguz Derin <[email protected]>
Fixes #4561 * Add shader- and pipeline-creation errors to smoothstep if `low >= high` * A line break to improve rendering of the formula
Uses this function for all GPUObjectBase-derived types to ensure that the internal `[[device]]` and `[[label]]` are properly set on creation.
* Remove [[internal]] concept * Addressing feedback * Update spec/index.bs Co-authored-by: Kai Ninomiya <[email protected]>
* Indicate timeline for majority of algorithms * Addressing feedback * Define 'any timeline' * Apply Kai's suggestions from code review * More feedback * Apply more Kai suggestions from code review --------- Co-authored-by: Kai Ninomiya <[email protected]>
Fixes #4623 * Add bit reinterpretation built-in functions to list of flushable functions
…s valid or not. (#4608)
Add the restriction that only a single aspect (stencil-only or depth-only) may be used specified per texture in a given draw.
* Add missing limits to the correspondence doc * Apply Kai's suggestions from code review --------- Co-authored-by: Kai Ninomiya <[email protected]>
…4641) * Add coordinate system diagrams and fix NDC to framebuffer equation * Apply Kai's suggestions from code review * Address more feedback * Made the diagrams responsive to light/dark mode * Apply more of Kai's suggestions from code review --------- Co-authored-by: Kai Ninomiya <[email protected]>
sample_mask and sample_index usage should genearte an error at shader module creation time, not pipleine creation time.
* Compat: disallow @interpolate(flat, first)
* Add a better definition for 'texture' * Addressing Corentin's feedback * Apply Kai's suggestion --------- Co-authored-by: Kai Ninomiya <[email protected]>
* Describe how comparison samplers interact with filtering * Apply Kai's suggestion * Fix inverted compare pass/fail values * Apply more of Kai's suggestions from code review --------- Co-authored-by: Kai Ninomiya <[email protected]>
* Deduped most image/buffer copy validation * Addressing more feedback * Make the algorithm work for source and destination textures * Move 'validating texture copy range' step under 'validating GPUImageCopyTexture'
* Make valid, invalid, and expired regular properties * Apply Kai's suggestions from code review * Fixing several TODOs * Rebase --------- Co-authored-by: Kai Ninomiya <[email protected]>
* add first and either to interpolate
* Switch "validating GPUVertexState" to algorithm style * Fix attribute validation
I believe these examples are left over from before WGSL had `for` loops. Now that WGSL has for loops these examples showing a `for` loop expressed with `loop` can directly reference `for` in WGSL. Fixed: #4996
Introduce float16-renderable and float32-renderable as optional features in Compatibility mode, and required in Core.
Add new limits for storage buffers and textures in the fragment stage.
* Add attributes * Normalize line endings * Trim trailing blankspaces * Make endings consistent
Now that GPUAdapter.featureLevel has landed in the core spec, remove the GPUAdapter.compatibilityMode IDL change from the proposal. Explain that setting GPUAdapter.featureLevel to "compatibility" enables Compatibility Mode validation. Change the GPUAdapter.isCompatibilityMode boolean to a "featureLevel" DOMString.
* texel buffer proposal: Update from review - Clarify Metal limits are in units of pixels - Add TODO to get info for pre-Apple silicon Metal GPUs - Add @teoxoy's question about uniform texel buffer as an open question Followup to #4912 * Update per review Resolve the 'uniform texel buffer' question: read-only views map to uniorm texel buffers. Add the question about needing another create-view method for these.
`Status: LD` is "Living Document", which makes sense but has broken CSS if used with `Group: webgpu`. If not used with `Group: webgpu` then the license is CC0 (Public Domain) which is probably fine but not really appropriate. `Status: UD` is "Unofficial Proposal Draft", which is not really what I would call this document, but it renders correctly :) Any other status causes an error message saying something like: "You used Status NOTE, but your Group (WEBGPU) is limited to the statuses CG-DRAFT, CG-FINAL, or UD." (Probably LD isn't supposed to be used either and that's why it's broken.) Documentation: https://speced.github.io/bikeshed/#metadata
New spec drafts need to be published as "Candidate Recommendation Drafts". This status requires a deadline date to be specified. That deadline could be set as a `W3C_BUILD_OVERRIDE` parameter in the workflow, but then it's inherent to the specs, so it seems good to have it defined in the source directly. (That information is not useful for the Editor's Draft in theory, but also needed in practice since Bikeshed replaces all macro texts in the boilerplate before it drops the bits that aren't needed). The update also adds a link to the test suite in WebGPU and renames "Tests" into "Test Suite" in WGSL (usual name across specs).
* API
* new feature: subgroups
* new properties for adapter info:
* subgroupMinSize
* subgroupMaxSize
* WGSL
* new enable: subgroups
* new built-in values
* subgroup_invocation_id
* subgroup_size
* subgroup and quad built-in functions
* new uniformity diagnostic subgroup_uniformity
This patch relaxes the check on the parameter `index` in `getBindGroupLayout()` by allowing `index < limits.maxBindGroups` instead of only limiting `index` must be less than the array size of the bind group layouts used in the creation of the pipeline.
…5048) Reinforces the fix to #4648 Co-authored-by: Mehmet Oguz Derin <[email protected]>
Should unbreak the build for now (but publishing will still be broken) speced/bikeshed#3012 We can make build-validate-publish required for PRs, and make publish-TR-webgpu and publish-TR-wgsl optional.
Here's the proposal in code
requestDevice({ requestedLimits }) {
// make a copy of the requested limits
const effectiveRequestedLimits = { ...requestedLimits };
// update the effective requested limits
effectiveRequestedLimits.maxStorageBuffersPerShaderStage = Math.max(
requestedLimits.maxStorageBuffersInFragmentStage,
requestedLimits.maxStorageBuffersInVertexStage,
requestedLimits.maxStorageBuffersPerShaderStage);
effectiveRequestedLimits.maxStorageTexturesPerShaderStage = Math.max(
requestedLimits.maxStorageTexturesInFragmentStage,
requestedLimits.maxStorageTexturesInVertexStage,
requestedLimits.maxStorageTexturesPerShaderStage);
// request the device using the effectiveRequestedLimits following the normal
// validation rules. If any requested limit exceeds the adapter's limit then
// throw an OperationError
// device creation succeeded
if (featureLevel === 'core') {
// this effectively leaves core validation the same as it has always been.
device.limits.maxStorageBuffersInFragmentStage = device.limit.maxStorageBuffersPerShaderStage;
device.limits.maxStorageBuffersInVertexStage = device.limit.maxStorageBufferPerShaderStage;
device.limits.maxStorageTexturesInFragmentStage = device.limit.maxStorageTexturesPerShaderStage;
device.limits.maxStorageTexturesInVertexStage = device.limit.maxStorageTexturesPerShaderStage;
}
With the limits set as above, apply the validation as specified in the proposal at createBindGroupLayout, createPipelineLayout (including "auto" layout creation)
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.
See Commits and Changes for more details.
Created by
pull[bot]
Can you help keep this open source service alive? 💖 Please sponsor : )