After merging your PR, Conbench analyzed the 6 benchmarking runs that have been run so far on merge-commit 8cc71ab.

Benchmarks with performance regressions

There were 5 possible performance regressions, according to the lookback z-score method.

Benchmarks with regressions:

• Commit Run on ursa-i9-9960x at 2023-11-22 16:59:04Z
  • file-write (Python) with compression=uncompressed, dataset=fanniemae_2016Q4, file_type=parquet, input_type=dataframe
  • dataset-selectivity (Python) with dataset=nyctaxi_multi_ipc_s3, selectivity=10%
  • file-write (Python) with compression=uncompressed, dataset=fanniemae_2016Q4, file_type=feather, input_type=dataframe
  • dataframe-to-table (R) with dataset=type_nested, language=R
  • dataframe-to-table (R) with dataset=type_dict, language=R

All benchmark runs analyzed:

• Commit Run on ursa-i9-9960x at 2023-11-22 16:59:04Z
• Commit Run on ursa-thinkcentre-m75q at 2023-11-22 19:19:05Z
• Commit Run on arm64-m6g-linux-compute at 2023-11-22 17:38:39Z
• Commit Run on ec2-t3-xlarge-us-east-2 at 2023-11-22 17:26:25Z
• Commit Run on ec2-m5-4xlarge-us-east-2 at 2023-11-22 17:03:02Z
• Commit Run on arm64-t4g-linux-compute at 2023-11-22 17:41:34Z

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 2023-11-22 19:19:05Z

  • MinMaxKernelDouble (C++) with params=1048576/100, source=cpp-micro, suite=arrow-acero-aggregate-benchmark
  • BM_BatchComputeHash (C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmark
  • BM_FindExistingHash (C++) with source=cpp-micro, suite=parquet-bloom-filter-benchmark
  • SumKernelInt8 (C++) with params=1048576/10000, source=cpp-micro, suite=arrow-acero-aggregate-benchmark
  • VarianceKernelFloat (C++) with params=1048576/1, source=cpp-micro, suite=arrow-acero-aggregate-benchmark
  • BM_BatchComputeHash (C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmark
  • BM_BatchComputeHash (C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmark
  • getNullCountBenchmark (Java) with source=java-micro, suite=arrow.vector.BitVectorHelperBenchmarks
  • BM_BatchComputeHash (C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmark

• Commit Run on arm64-m6g-linux-compute at 2023-11-22 17:38:39Z

  • ConstructFromShortStdVector (C++) with params=<STATIC_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmark
  • CopyShortVector (C++) with params=<SMALL_VECTOR(std::shared_ptr)>, source=cpp-micro, suite=arrow-small-vector-benchmark
  • CopyShortVector (C++) with params=<SMALL_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmark
  • CopyShortVector (C++) with params=<SMALL_VECTOR(int)>, source=cpp-micro, suite=arrow-small-vector-benchmark
  • BitRunReaderLinear (C++) with params=-1, source=cpp-micro, suite=arrow-bit-util-benchmark