Remove lifetime constraints from wgpu::ComputePass methods
#5570
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
Wumpf
force-pushed
the
lifetime-free-computepass
branch
from
c7655a4
to
b976ff1
Compare
This was referenced Apr 21, 2024
Separate out ComputeCommand id->arc resolve (a step towards no lifetimes on
wgpu::ComputePass)
#5432
Merged
…rectly on recording TODO: * bind groups don't work because the Binder gets an id only * wgpu level error handling is missing
…eeds to lock bind_group storage
… DynComputePass on wgpu side
Wumpf
force-pushed
the
lifetime-free-computepass
branch
from
b976ff1
to
0d470a9
Compare
Wumpf
commented
Apr 23, 2024
Wumpf
commented
Apr 23, 2024
6 tasks
6 tasks
cwfitzgerald
approved these changes
May 14, 2024
Me too! I really want to add a compute pass benchmark and backport it now that you started that work |
jimblandy
pushed a commit
to jimblandy/wgpu
that referenced
this pull request
May 23, 2024
…5570) * basic test setup * remove lifetime and drop resources on test - test fails now just as expected * compute pass recording is now hub dependent (needs gfx_select) * compute pass recording now bumps reference count of uses resources directly on recording TODO: * bind groups don't work because the Binder gets an id only * wgpu level error handling is missing * simplify compute pass state flush, compute pass execution no longer needs to lock bind_group storage * wgpu sided error handling * make ComputePass hal dependent, removing command cast hack. Introduce DynComputePass on wgpu side * remove stray repr(C) * changelog entry * fix deno issues -> move DynComputePass into wgc * split out resources setup from test
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.
Connections
wgpu::ComputePass) #5432wgpu::ComputePass) #5432 lands as this contains all the changes again.RenderPassmake it difficult to use. #1453Part of a series towards lifetime removal on compute passes:
wgpu::ComputePass) #5432wgpu::ComputePassmethods #5570ComputePassto its parent command encoder #5620Description
Removes the need for lifetime dependency on any resource passed into a ComputePass.
Lifetime constraints to the originating encoder & queries used during creation still apply for the moment (to be addressed in a follow-up.
Directly builds on #5432 by moving the creation of
ArcComputeCommandinto compute pass methods.To accomplish this I had to add a HalApi generic dependency on
ComputePass. Since I didn't want to introduce a new object to the id/objecttracking system, I instead decided to create aDynComputePassonwgpu/wgpu_core.rsside. Hopefully we can simplify this as part of the generic refactor (#5124). Note that the new lifetime dependency made boxing of the wgpu-core compute pass is unavoidable at this point andDynComputePassallows us to merge the otherwise requiredgfx_select& type cast into a single trait/virtual call.An alternative I expolored was to do the type erasure internally on
BasePass<ArcComputeCommand<A>>but I found this quite unwieldy, also this wouldn't cover timestamp writes which are intentionally left unaddressed in this PR.Testing
Added a new test that ensures that resources passed into a ComputePass can be dropped before executing the compute pass.
Proved that the test would crash before these changes and passes with them.
Checklist
cargo fmt.cargo clippy. If applicable, add:--target wasm32-unknown-unknown--target wasm32-unknown-emscriptencargo xtask testto run tests.CHANGELOG.md. See simple instructions inside file.