After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit 5b68eca.

Benchmarks with performance regressions

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

All benchmark runs analyzed:

• Commit Run on arm64-t4g-2xlarge-linux at 2024-10-30 23:34:52Z
• Commit Run on amd64-c6a-4xlarge-linux at 2024-10-30 23:15:52Z
• Commit Run on amd64-m5-4xlarge-linux at 2024-10-30 23:30:05Z

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 arm64-t4g-2xlarge-linux at 2024-10-30 23:34:52Z

  • ProjectionOverhead (C++) with params=zero_copy_expression/batch_size:10000/real_time, source=cpp-micro, suite=arrow-acero-project-benchmark

• Commit Run on amd64-c6a-4xlarge-linux at 2024-10-30 23:15:52Z

  • ConstructFromShortStdVector (C++) with params=<SMALL_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmark
  • ConstructFromEmptyStdVector (C++) with params=<SMALL_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmark
  • BenchmarkTemporalBinary (C++) with params=<MonthDayNanoBetween, zoned>/524288/0, source=cpp-micro, suite=arrow-compute-scalar-temporal-benchmark