After merging your PR, Conbench analyzed the 2 benchmarking runs that have been run so far on merge-commit e5c8ccd.

Benchmarks with performance regressions

There were 0 possible performance regressions, according to the lookback z-score method.

All benchmark runs analyzed:

• Commit Run on test-mac-arm at 2025-10-07 20:08:15Z
• Commit Run on arm64-t4g-2xlarge-linux at 2025-10-07 17:14:22Z

Unstable benchmarks with performance regressions

The following benchmark results indicate a possible performance regression, but are known to sometimes produce false positives when applying the lookback z-score analysis.

• Commit Run on test-mac-arm at 2025-10-07 20:08:15Z

  • FilterOverhead (C++) with params=not_null_and_is_true_singlepass_benchmark/batch_size:10000/null_prob:0/bool_true_prob:25/real_time, source=cpp-micro, suite=arrow-acero-filter-benchmark
  • FilterOverhead (C++) with params=not_null_and_is_true_singlepass_benchmark/batch_size:1000/null_prob:25/bool_true_prob:33/real_time, source=cpp-micro, suite=arrow-acero-filter-benchmark

• Commit Run on arm64-t4g-2xlarge-linux at 2025-10-07 17:14:22Z

  • TouchArea (C++) with params=size:65536/real_time/threads:16, source=cpp-micro, suite=arrow-memory-pool-benchmark
  • BM_DecodeArrowBooleanRle (C++) with params=DecodeArrowNonNull/32768, source=cpp-micro, suite=parquet-encoding-benchmark
  • BM_ByteStreamSplitEncode_Int16_Neon (C++) with params=1024, source=cpp-micro, suite=parquet-encoding-benchmark
  • TouchArea (C++) with params=size:65536/real_time/threads:8, source=cpp-micro, suite=arrow-memory-pool-benchmark