GH-48277: [C++][Parquet] unpack with shuffle algorithm#47994
Merged
conbench-apache-arrow / Conbench performance report
failed
Feb 26, 2026 in 0s
Found 32 regressions
Thanks for your patience. Conbench analyzed the 3 benchmarking runs that have been run so far on PR commit 8f18e86.
Benchmarks with performance regressions
There were 32 possible performance regressions, according to the lookback z-score method.
Benchmarks with regressions:
-
Pull Request Run on
amd64-c6a-4xlarge-linuxat 2026-02-26 19:51:53ZBM_UnpackUint64(C++) with params=DynamicAligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=DynamicUnaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint16(C++) with params=DynamicUnaligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint32(C++) with params=DynamicUnaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint32(C++) with params=Avx2Unaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=Avx2Unaligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=Avx2Unaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint16(C++) with params=Avx2Unaligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=ScalarUnaligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=DynamicUnaligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=DynamicAligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint32(C++) with params=DynamicAligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_ByteStreamSplitEncode_Int16_Scalar(C++) with params=65536, source=cpp-micro, suite=parquet-encoding-benchmarkBM_UnpackUint16(C++) with params=DynamicAligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmark
-
Pull Request Run on
amd64-m5-4xlarge-linuxat 2026-02-26 20:05:22ZBM_UnpackUint64(C++) with params=Avx2Unaligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=Avx2Unaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint16(C++) with params=Avx2Unaligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmark
-
Pull Request Run on
arm64-t4g-2xlarge-linuxat 2026-02-26 20:15:57ZBM_DeltaBitPackingDecode_Int64_Wide(C++) with params=4096, source=cpp-micro, suite=parquet-encoding-benchmarkBM_UnpackUint64(C++) with params=DynamicAligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_DeltaBitPackingDecode_Int64_Wide(C++) with params=1024, source=cpp-micro, suite=parquet-encoding-benchmarkBM_UnpackUint32(C++) with params=NeonUnaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_DeltaBitPackingDecode_Int64_Wide(C++) with params=32768, source=cpp-micro, suite=parquet-encoding-benchmarkBM_UnpackUint64(C++) with params=NeonUnaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=DynamicUnaligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint32(C++) with params=DynamicUnaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=DynamicUnaligned/20/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint16(C++) with params=NeonUnaligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=DynamicAligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint16(C++) with params=DynamicUnaligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint64(C++) with params=NeonUnaligned/47/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_UnpackUint16(C++) with params=DynamicAligned/13/64, source=cpp-micro, suite=arrow-bpacking-benchmarkBM_DeltaBitPackingDecode_Int64_Wide(C++) with params=65536, source=cpp-micro, suite=parquet-encoding-benchmark
All benchmark runs analyzed:
- Pull Request Run on
amd64-c6a-4xlarge-linuxat 2026-02-26 19:51:53Z - Pull Request Run on
amd64-m5-4xlarge-linuxat 2026-02-26 20:05:22Z - Pull Request Run on
arm64-t4g-2xlarge-linuxat 2026-02-26 20:15:57Z
Details
This report was generated using the lookback z-score method with a z-score threshold of 30.0.
Loading