vortex-data/vortex - CodSpeed

vortex

Blog Docs Changelog

Performance History

Latest Results

more Signed-off-by: Robert Kruszewski <github@robertk.io>

refactor/parent-ref-stack-allocated

5 hours ago

lock Signed-off-by: Robert Kruszewski <github@robertk.io>

refactor/parent-ref-stack-allocated

7 hours ago

perf(array): drop redundant FlatBuffer verifier in SerializedArray::buffer_lengths `SerializedArray::buffer_lengths()` was calling `root::<fba::Array>()` on every invocation, which re-runs the full FlatBuffer verifier across the array tree. The buffer is already verified once at construction time by `validate_array_tree` (line 527), and the rest of `SerializedArray` already exploits this invariant - see `from_flatbuffer_and_segment_with_overrides` at line 614, which uses `fba::root_as_array_unchecked` with the same safety justification. Switch `buffer_lengths` to `fba::root_as_array_unchecked` for consistency. Measured on the same workload with the new buffer_lengths bench: n_fields fixed legacy(root::<Array>) speedup 1 26 ns 237 ns 9x 8 35 ns 772 ns 22x 32 59 ns 2.56 us 43x 100 140 ns 7.75 us 55x 1000 1.13 us 73.7 us 65x Post-fix cost grows with n_fields purely from the Vec<usize> allocation + iteration to extract buffer descriptor lengths; the verifier overhead is gone. Signed-off-by: Claude <noreply@anthropic.com>

claude/flatbuffers-memory-safety-XKbWQ

8 hours ago

simpler Signed-off-by: Robert Kruszewski <github@robertk.io>

refactor/parent-ref-stack-allocated

9 hours ago

locks Signed-off-by: Robert Kruszewski <github@robertk.io>

refactor/parent-ref-stack-allocated

9 hours ago

u Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>

claude/bitpack-compare-speedup-KGPS3

11 hours ago

Merge remote-tracking branch 'origin/develop' into claude/bitpack-compare-speedup-KGPS3 Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk> # Conflicts: # encodings/fastlanes/Cargo.toml

claude/bitpack-compare-speedup-KGPS3

11 hours ago

bench: bit-packed compare-constant baseline (#8012) ## Summary - Adds a divan benchmark `bitpack_compare` in `vortex-fastlanes` that compares an `Operator::Eq` / `Operator::Lt` against an out-of-range constant on a `BitPackedData` array vs. an explicit "decompress, then Arrow compare" baseline that materialises the unpacked `PrimitiveArray` first. - The constant is chosen as `1 << BW`, just past the packable range, so a future compare-constant kernel can recognise it and short-circuit. Today both arms decompress; this PR establishes a baseline for that follow-up to land against. - Grid sized for fast runs: `len ∈ {1024, 65536}`, `bit_width ∈ {4, 16}`, Eq + Lt. The follow-up optimization (out-of-range fast path on `BitPacked`, plus the `bitpack_constant` analytical encoder) is in #PR2-PLACEHOLDER, stacked on this branch. Splitting the bench out lets the speedup PR show concrete numbers against this measured baseline. ## Test plan - [x] `cargo check -p vortex-fastlanes --benches` - [ ] `cargo bench -p vortex-fastlanes --bench bitpack_compare` records the slow baseline numbers prior to the follow-up landing 🤖 Generated with [Claude Code](https://claude.com/claude-code) Signed-off-by: Joe Isaacs <joe.isaacs@live.co.uk>

develop

11 hours ago

Latest Branches

+7%

Allow running reduce_parent operations on stack allocated parents#7751

5 hours ago

3f45a2a

refactor/parent-ref-stack-allocated

+3%

[claude] bench: measure FlatBuffers verification cost for Layout and Array#8014

20 hours ago

172e0c7

claude/flatbuffers-memory-safety-XKbWQ

×2.6

fastlanes: bit-packed compare-constant fast path + bitpack_constant kernel#8013

11 hours ago

3b1b8cf

claude/bitpack-compare-speedup-KGPS3

© 2026 CodSpeed Technology

Home Terms Privacy Docs