All pull requests

Pull requests

• Speed up scalar construction, stats reads and aggregate partial merging

  changelog/break
  #10326·robert3005 opened Oct 6, 2026Contributor·

  ·
  1

• perf(layout): use fearless_simd for bloom filter block masks

  #10325·joseph-isaacs opened Oct 6, 2026Contributor·

  ·
  1

• perf(array): use fearless_simd for AVX2 compress and is_constant

  #10324·joseph-isaacs opened Oct 6, 2026Contributor·

  ·
  1

• perf(buffer): use fearless_simd for x86 popcount

  #10323·joseph-isaacs opened Oct 6, 2026Contributor·

  ·
  1

• [codex] Experiment: Fearless SIMD byte compression regression

  #10321·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• [codex] Experiment: Fearless SIMD boolean packing

  #10320·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• [codex] Speed up ARM64 bitmap rank-select with Fearless SIMD

  #10319·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• [codex] Accelerate small boolean dictionary takes on aarch64

  #10318·joseph-isaacs opened Oct 5, 2026Contributor·

  ·

• perf(array): extend small-table take SIMD

  #10317·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  2

• perf(runend): coalesce repeated boolean predicate runs

  #10316·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  25

• Add bounded decimal dispatch macros and migrate kernels

  #10314·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• Increase ArraySlots inline capacity to five

  changelog/skip
  #10312·mhk197 opened Oct 5, 2026Contributor·

  ·
  28

• feat(tui): add encoding kernel coverage diagnostics

  #10310·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• [reminder] chore: prepare DataFusion 56 upgrade

  changelog/break
  #10309·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• bench(runend): compare adaptive kernels on AVX2 and NEON

  #10308·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• Bound decimal storage by precision

  #10305·joseph-isaacs opened Oct 5, 2026Contributor·

  ·

• [EXPERIMENTAL] perf: make checked lane chunking opt-in

  changelog/performance
  #10304·connortsui20 opened Oct 5, 2026Member·

  ·
  1

• Add an opt-in V2 scan builder and execution path

  changelog/chore
  #10302·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  1

• perf(array): avoid errors when comparison kernels decline arithmetic

  changelog/performance
  #10301·connortsui20 opened Oct 5, 2026Member·

  ·
  1

• perf(compute): tile checked lane kernels consistently

  changelog/performance
  #10299·connortsui20 opened Oct 5, 2026Member·

  ·
  12

• perf(runend): decode primitive runs per segment with branch-free head stores

  #10296·joseph-isaacs opened Oct 5, 2026Contributor·

  ·
  12

• fix(buffer): don't create &mut [u64] over uninitialized memory in collect_words_in

  action/bench-sqlchangelog/fix
  #10294·jianhe25 opened Oct 5, 2026Contributor·

  ·
  1

• fix(buffer): require both halves of a Zip to be TrustedLen

  changelog/fix
  #10293·jianhe25 opened Oct 5, 2026Contributor·

  ·
  2

• fix(buffer): validate offsets in Buffer::into_arrow_offset_buffer

  #10292·jianhe25 opened Oct 5, 2026Contributor·

  ·
  1

• fix(ipc): fill the buffer on partial reads in the sync message reader

  #10284·jackylee-ch opened Oct 4, 2026Contributor·

  ·