Replace ToCanonical trait usages with execute#8609
Merged
Conversation
Merging this PR will degrade performance by 10.5%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ❌ | Simulation | extend_from_array_zctl[(1000, 8)] |
239.6 µs | 409.6 µs | -41.51% |
| ❌ | Simulation | extend_from_array_non_zctl_overlapping[(1000, 8)] |
272.2 µs | 444.7 µs | -38.79% |
| ❌ | Simulation | extend_from_array_non_zctl_overlapping[(1000, 32)] |
737.7 µs | 909.8 µs | -18.92% |
| ❌ | Simulation | compact_sliced[(4096, 90)] |
750 ns | 896.4 ns | -16.33% |
| ❌ | Simulation | extend_from_array_zctl[(1000, 64)] |
1 ms | 1.2 ms | -13.98% |
| ❌ | Simulation | compact_sliced[(16384, 90)] |
813.3 ns | 930.6 ns | -12.6% |
| ❌ | Simulation | sum_i32_nullable_all_valid |
37.2 µs | 42.1 µs | -11.76% |
| ❌ | Simulation | extend_from_array_zctl[(10000, 8)] |
1.9 ms | 2.2 ms | -10.8% |
| ❌ | Simulation | encode_varbin[(1000, 2)] |
141.5 µs | 158.4 µs | -10.67% |
| ❌ | Simulation | encode_varbin[(1000, 4)] |
142.4 µs | 159.2 µs | -10.6% |
| ❌ | Simulation | extend_from_array_non_zctl_overlapping[(10000, 8)] |
2.2 ms | 2.5 ms | -10.57% |
| ❌ | Simulation | encode_varbin[(1000, 8)] |
143.9 µs | 160.7 µs | -10.44% |
| ❌ | Simulation | chunked_varbinview_opt_canonical_into[(100, 100)] |
305.7 µs | 340.5 µs | -10.22% |
| ❌ | Simulation | encode_varbin[(1000, 32)] |
147.8 µs | 164.5 µs | -10.17% |
| ⚡ | Simulation | bench_many_codes_few_values[1024] |
538.9 µs | 363.9 µs | +48.07% |
| ⚡ | Simulation | chunked_varbinview_canonical_into[(1000, 10)] |
191.1 µs | 154.7 µs | +23.48% |
| ⚡ | Simulation | bitwise_not_vortex_buffer_mut[128] |
273.6 ns | 244.4 ns | +11.93% |
Tip
Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.
Comparing claude/nice-hypatia-g68zgt (f3bf498) with develop (4959939)
Footnotes
-
42 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩
This was referenced Jun 27, 2026
robert3005
force-pushed
the
claude/nice-hypatia-g68zgt
branch
from
5e975ca to
f8f6a90
Compare
lwwmanning
reviewed
Jun 27, 2026
robert3005
force-pushed
the
claude/nice-hypatia-g68zgt
branch
2 times, most recently
from
048b19b to
947a966
Compare
robert3005
added a commit
that referenced
this pull request
Jun 27, 2026
Rebased onto develop and made standalone (no longer stacked on #8609). Instead of deprecating the hidden global session, expose it as `vortex_array::legacy_session()` and add it to the central `clippy.toml` `disallowed-methods` list (denied via `clippy::all`). New call sites are now rejected by clippy; every current invocation is explicitly allowlisted with `#[allow(clippy::disallowed_methods)]` on its enclosing function. - `LEGACY_SESSION` static -> `legacy_session() -> &'static VortexSession` (the `LazyLock` is now a private static inside the accessor). - All call sites renamed and allowlisted across vortex-array, the encodings, vortex-ffi, vortex-cuda, vortex-cxx, vortex-layout, vortex-python and vortex-web. develop has more call sites than #8609 (incl. test/builder code and crates #8609 had refactored away), all covered. Verified with: - cargo clippy -p vortex-array -p vortex-alp -p vortex-runend -p vortex-fsst -p vortex-fastlanes -p vortex-zstd -p vortex-ffi -p vortex-layout -p vortex-web-wasm -p vortex-cxx -p vortex-python --all-targets --all-features (0 disallowed-method hits, 0 warnings) - cargo +nightly fmt --check (touched crates) vortex-cuda could not be compiled here (CUDA toolchain unavailable); its two call sites are allowlisted following the same pattern. Signed-off-by: Robert Kruszewski <robert@spiraldb.com> Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018E5DywGCpq26ndoMHd1vyg
robert3005
force-pushed
the
claude/nice-hypatia-g68zgt
branch
2 times, most recently
from
bc088d1 to
d2f276a
Compare
robert3005
added a commit
that referenced
this pull request
Jul 1, 2026
Rebased onto develop and made standalone (no longer stacked on #8609). Instead of deprecating the hidden global session, expose it as `vortex_array::legacy_session()` and add it to the central `clippy.toml` `disallowed-methods` list (denied via `clippy::all`). New call sites are now rejected by clippy; every current invocation is explicitly allowlisted with `#[allow(clippy::disallowed_methods)]` on its enclosing function. - `LEGACY_SESSION` static -> `legacy_session() -> &'static VortexSession` (the `LazyLock` is now a private static inside the accessor). - All call sites renamed and allowlisted across vortex-array, the encodings, vortex-ffi, vortex-cuda, vortex-cxx, vortex-layout, vortex-python and vortex-web. develop has more call sites than #8609 (incl. test/builder code and crates #8609 had refactored away), all covered. Verified with: - cargo clippy -p vortex-array -p vortex-alp -p vortex-runend -p vortex-fsst -p vortex-fastlanes -p vortex-zstd -p vortex-ffi -p vortex-layout -p vortex-web-wasm -p vortex-cxx -p vortex-python --all-targets --all-features (0 disallowed-method hits, 0 warnings) - cargo +nightly fmt --check (touched crates) vortex-cuda could not be compiled here (CUDA toolchain unavailable); its two call sites are allowlisted following the same pattern. Signed-off-by: Robert Kruszewski <robert@spiraldb.com> Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018E5DywGCpq26ndoMHd1vyg Signed-off-by: Robert Kruszewski <github@robertk.io>
robert3005
added
changelog/break
Contributor
|
@codex please review this change |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: de998866cb
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
robert3005
force-pushed
the
claude/nice-hypatia-g68zgt
branch
from
08472bb to
0871721
Compare
robert3005
force-pushed
the
claude/nice-hypatia-g68zgt
branch
from
0871721 to
53ed496
Compare
This also removes ArrayBuilder::extend_from_array and moves that logic in to ArrayRef::append_to_builder Plumb ExecutionCtx through more methods instead of creating fresh one. Signed-off-by: Robert Kruszewski <github@robertk.io>
Replace the ListBuilder/ListViewBuilder downcast-enumeration macros with two dyn-dispatched ArrayBuilder methods, append_list_array and append_listview_array, taking ArrayViews. The builders resolve their own offset/size generics, so every IntegerPType (signed included) works, and either list encoding can append into either list builder. List-encoded sources now append directly from the ListArray layout without canonicalizing to ListViewArray first. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Signed-off-by: Robert Kruszewski <github@robertk.io>
robert3005
force-pushed
the
claude/nice-hypatia-g68zgt
branch
from
53ed496 to
f3bf498
Compare
Contributor
|
@codex review again pls |
|
Codex Review: Didn't find any major issues. Nice work! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Member
|
Seems like this improves polarsignals benchmark by a huge amount? Is that real? |
Contributor
Author
|
I think it is real. We have sometimes not delegated through VTable::append_to_builder which this pr fixes. I didn't know about this bug until we hit it elsewhere |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
This also removes ArrayBuilder::extend_from_array and moves that logic
in to ArrayRef::append_to_builder
Plumb ExecutionCtx through more methods instead of creating fresh one.
Fixes: #3235
Signed-off-by: Robert Kruszewski github@robertk.io