After merging your PR, Conbench analyzed the 6 benchmarking runs that have been run so far on merge-commit 7c0451c.

Benchmarks with performance regressions

There were 2 possible performance regressions, according to the lookback z-score method.

Benchmarks with regressions:

• Commit Run on arm64-t4g-linux-compute at 2023-07-26 20:59:41Z

  • tpch (R) with engine=arrow, format=native, language=R, memory_map=False, query_id=TPCH-22, scale_factor=1

• Commit Run on ursa-i9-9960x at 2023-08-02 20:16:22Z

  • dataframe-to-table (Python) with dataset=type_integers

All benchmark runs analyzed:

• Commit Run on ec2-t3-xlarge-us-east-2 at 2023-07-26 19:39:27Z
• Commit Run on arm64-t4g-linux-compute at 2023-07-26 20:59:41Z
• Commit Run on ec2-m5-4xlarge-us-east-2 at 2023-07-26 22:41:02Z
• Commit Run on arm64-m6g-linux-compute at 2023-07-27 00:53:33Z
• Commit Run on ursa-thinkcentre-m75q at 2023-07-28 12:05:19Z
• Commit Run on ursa-i9-9960x at 2023-08-02 20:16:22Z

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 ursa-thinkcentre-m75q at 2023-07-28 12:05:19Z
  • BitmapEqualsWithOffset (C++) with params=8192, source=cpp-micro, suite=arrow-bit-util-benchmark
  • BM_DeltaBitPackingEncode_Int32_Narrow (C++) with params=32768, source=cpp-micro, suite=parquet-encoding-benchmark