GH-47012: [C++][Parquet] Reserve values correctly when reading BYTE_A…
conbench-apache-arrow / Conbench performance report
failed
Jul 8, 2025 in 0s
Found 128 regressions
Thanks for your patience. Conbench analyzed the 4 benchmarking runs that have been run so far on PR commit 6067fe6.
Benchmarks with performance regressions
There were 128 possible performance regressions, according to the lookback z-score method.
Benchmarks with regressions:
-
Pull Request Run on
amd64-c6a-4xlarge-linuxat 2025-07-08 10:39:44ZBM_PlainEncodingSpacedBoolean(C++) with params=32768/100, source=cpp-micro, suite=parquet-encoding-benchmarkBM_BatchComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_DictDecodingInt64_literals(C++) with params=4096, source=cpp-micro, suite=parquet-encoding-benchmarkBM_ComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_PlainEncodingSpacedBoolean(C++) with params=32768/1, source=cpp-micro, suite=parquet-encoding-benchmarkBM_RleDecoding(C++) with params=65536/16, source=cpp-micro, suite=parquet-column-io-benchmarkListSliceInt64ListViewWithStop(C++) with params=32768/1, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkBM_ComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkInt8StridedTensorConversionFixture(C++) with params=/ConvertToSparseCSFTensorInt64, source=cpp-micro, suite=arrow-tensor-conversion-benchmarkBM_RleDecoding(C++) with params=4096/1, source=cpp-micro, suite=parquet-column-io-benchmarkBM_ComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_RleDecoding(C++) with params=32768/16, source=cpp-micro, suite=parquet-column-io-benchmarkFloatColumnMajorTensorConversionFixture(C++) with params=/ConvertToSparseCSFTensorInt64, source=cpp-micro, suite=arrow-tensor-conversion-benchmarkBM_RleDecoding(C++) with params=4096/16, source=cpp-micro, suite=parquet-column-io-benchmarkBM_RleEncoding(C++) with params=32768/1, source=cpp-micro, suite=parquet-column-io-benchmarkFloatRowMajorTensorConversionFixture(C++) with params=/ConvertToSparseCSFTensorInt64, source=cpp-micro, suite=arrow-tensor-conversion-benchmarkInt8RowMajorTensorConversionFixture(C++) with params=/ConvertToSparseCSFTensorInt64, source=cpp-micro, suite=arrow-tensor-conversion-benchmarkIntegerFormatting(C++) with params=, source=cpp-micro, suite=arrow-value-parsing-benchmarkBM_RleEncoding(C++) with params=1024/1, source=cpp-micro, suite=parquet-column-io-benchmarkBM_RleEncoding(C++) with params=65536/1, source=cpp-micro, suite=parquet-column-io-benchmarkBM_RleEncoding(C++) with params=4096/1, source=cpp-micro, suite=parquet-column-io-benchmarkBM_FindExistingHash(C++) with source=cpp-micro, suite=parquet-bloom-filter-benchmarkInt8ColumnMajorTensorConversionFixture(C++) with params=/ConvertToSparseCSFTensorInt64, source=cpp-micro, suite=arrow-tensor-conversion-benchmarkMemoizeLruCachedThreadUnsafe(C++) with params=64/16, source=cpp-micro, suite=arrow-cache-benchmarkBM_RleDecoding(C++) with params=32768/1, source=cpp-micro, suite=parquet-column-io-benchmark
-
Pull Request Run on
arm64-t4g-2xlarge-linuxat 2025-07-08 12:09:38ZShortVectorInsertAtEnd(C++) with params=<STD_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkCheckIndexBoundsUInt32(C++) with params=1048576/0, source=cpp-micro, suite=arrow-int-util-benchmarkConstructFromEmptyStdVector(C++) with params=<SMALL_VECTOR(std::shared_ptr)>, source=cpp-micro, suite=arrow-small-vector-benchmarkCheckIndexBoundsInt32(C++) with params=1048576/1, source=cpp-micro, suite=arrow-int-util-benchmarkLongVectorInsertAtEnd(C++) with params=<SMALL_VECTOR(int)>, source=cpp-micro, suite=arrow-small-vector-benchmarkRecordReaderReadRecords(C++) with params=Repetition:1/BatchSize:1000/ReadDense:1, source=cpp-micro, suite=parquet-column-reader-benchmarkCopyShortVector(C++) with params=<STATIC_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkConstructFromShortStdVector(C++) with params=<SMALL_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkConstructFromShortStdVector(C++) with params=<STATIC_VECTOR(int)>, source=cpp-micro, suite=arrow-small-vector-benchmarkAllocateDeallocate(C++) with params=/size:1048576/real_time/threads:1, source=cpp-micro, suite=arrow-memory-pool-benchmarkCheckIndexBoundsUInt32(C++) with params=65536/0, source=cpp-micro, suite=arrow-int-util-benchmarkMoveEmptyVector(C++) with params=<STATIC_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkShortVectorInsertAtEnd(C++) with params=<STD_VECTOR(int)>, source=cpp-micro, suite=arrow-small-vector-benchmarkCheckIndexBoundsUInt32(C++) with params=65536/1000, source=cpp-micro, suite=arrow-int-util-benchmarkCheckIndexBoundsUInt32(C++) with params=1048576/1000, source=cpp-micro, suite=arrow-int-util-benchmarkCheckIndexBoundsUInt32(C++) with params=65536/1, source=cpp-micro, suite=arrow-int-util-benchmarkCheckIndexBoundsInt32(C++) with params=65536/2, source=cpp-micro, suite=arrow-int-util-benchmarkConstructFromShortStdVector(C++) with params=<STD_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkCopyEmptyVector(C++) with params=<STD_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkCopyShortVector(C++) with params=<STATIC_VECTOR(std::shared_ptr)>, source=cpp-micro, suite=arrow-small-vector-benchmarkMoveShortVector(C++) with params=<STATIC_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkMoveShortVector(C++) with params=<STATIC_VECTOR(std::shared_ptr)>, source=cpp-micro, suite=arrow-small-vector-benchmarkCheckIndexBoundsInt32(C++) with params=65536/1, source=cpp-micro, suite=arrow-int-util-benchmarkConstructFromEmptyStdVector(C++) with params=<SMALL_VECTOR(int)>, source=cpp-micro, suite=arrow-small-vector-benchmark
-
Pull Request Run on
amd64-m5-4xlarge-linuxat 2025-07-08 11:24:41ZAllocateTouchDeallocate(C++) with params=/size:1048576/real_time/threads:2, source=cpp-micro, suite=arrow-memory-pool-benchmarkIndexInInt64SmallSet(C++) with params=4, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkBM_DecodeArrowBooleanPlain(C++) with params=DecodeArrow/1024, source=cpp-micro, suite=parquet-encoding-benchmarkShortVectorInsertAtEnd(C++) with params=<SMALL_VECTOR(std::string)>, source=cpp-micro, suite=arrow-small-vector-benchmarkBM_BatchComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkListSliceStringListViewWithStepAndStop(C++) with params=32768/100, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkArrayRangeEqualsDenseUnion(C++) with params=32768/10000, source=cpp-micro, suite=arrow-compare-benchmarkRecordReaderReadAndSkipRecords(C++) with params=Repetition:1/BatchSize:10000/LevelsPerPage:1000000, source=cpp-micro, suite=parquet-column-reader-benchmarkErrorSchemeResultNoInline(C++) with source=cpp-micro, suite=arrow-type-benchmarkIndexInInt64SmallSet(C++) with params=16, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkIndexInInt32SmallSet(C++) with params=64, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkColumnReaderReadBatchInt32(C++) with params=Repetition:1/BatchSize:1000, source=cpp-micro, suite=parquet-column-reader-benchmarkCoalesceBench64(C++) with params=5, source=cpp-micro, suite=arrow-compute-scalar-if-else-benchmarkArrayRangeEqualsDenseUnion(C++) with params=32768/2, source=cpp-micro, suite=arrow-compare-benchmarkBM_DeltaDecodingByteArray(C++) with params=max-string-length:1024/batch-size:2048/prefixed-percent:99, source=cpp-micro, suite=parquet-encoding-benchmarkBM_ReadListOfStructColumn(C++) with params=99, source=cpp-micro, suite=parquet-arrow-reader-writer-benchmarkIsInInt32SmallSet(C++) with params=16, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkIndexInInt64SmallSet(C++) with params=64, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkIndexInInt64SmallSet(C++) with params=2, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkArrayRangeEqualsDenseUnion(C++) with params=32768/0, source=cpp-micro, suite=arrow-compare-benchmarkBM_ReadInt64Column(C++) with params=<Repetition::REQUIRED, Compression::ZSTD, Encoding::BYTE_STREAM_SPLIT>/1024/1024, source=cpp-micro, suite=parquet-column-io-benchmarkBM_ReadInt64Column(C++) with params=Repetition::REQUIRED/1024/16, source=cpp-micro, suite=parquet-column-io-benchmarkRecordReaderSkipRecords(C++) with params=Repetition:1/BatchSize:1000, source=cpp-micro, suite=parquet-column-reader-benchmarkBM_RleDecoding(C++) with params=32768/16, source=cpp-micro, suite=parquet-column-io-benchmarkBM_RleDecoding(C++) with params=4096/8, source=cpp-micro, suite=parquet-column-io-benchmarkBM_BatchComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_DecodeArrowBooleanPlain(C++) with params=DecodeArrowWithNull/num_values:65536/null_in_ten_thousand:10000, source=cpp-micro, suite=parquet-encoding-benchmarkArrayRangeEqualsFixedSizeBinary(C++) with params=32768/2, source=cpp-micro, suite=arrow-compare-benchmarkListSliceInt64ListViewWithStop(C++) with params=32768/1, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkIsInInt64SmallSet(C++) with params=4, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkIndexInInt32SmallSet(C++) with params=16, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkBenchmarkBitmapVisitBitsetAnd(C++) with params=32768/1, source=cpp-micro, suite=arrow-bit-util-benchmarkIsInInt32SmallSet(C++) with params=2, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkTakeChunkedChunkedStringRandomIndicesWithNulls(C++) with params=1048576/0, source=cpp-micro, suite=arrow-compute-vector-selection-benchmarkIsInInt16SmallSet(C++) with params=64, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkIsInInt64SmallSet(C++) with params=16, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkArrayRangeEqualsDenseUnion(C++) with params=32768/10, source=cpp-micro, suite=arrow-compare-benchmarkArrayRangeEqualsDenseUnion(C++) with params=32768/100, source=cpp-micro, suite=arrow-compare-benchmarkAllocateTouchDeallocate(C++) with params=/size:65536/real_time/threads:16, source=cpp-micro, suite=arrow-memory-pool-benchmarkIndexInInt16SmallSet(C++) with params=64, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkColumnReaderSkipInt32(C++) with params=Repetition:2/BatchSize:1000, source=cpp-micro, suite=parquet-column-reader-benchmarkBM_DefinitionLevelsToBitmapRepeatedAllPresent(C++) with source=cpp-micro, suite=parquet-level-conversion-benchmarkBM_BatchComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_ReadInt64Column(C++) with params=<Repetition::OPTIONAL, Compression::ZSTD, Encoding::BYTE_STREAM_SPLIT>/1024/1024, source=cpp-micro, suite=parquet-column-io-benchmarkBM_DefinitionLevelsToBitmapRepeatedAllMissing(C++) with source=cpp-micro, suite=parquet-level-conversion-benchmarkIsInInt64SmallSet(C++) with params=64, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkListSliceInt64ListWithStepAndStop(C++) with params=32768/10, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkListSliceInt64ListWithStepAndStop(C++) with params=32768/10000, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkBM_RleDecoding(C++) with params=32768/8, source=cpp-micro, suite=parquet-column-io-benchmarkIsInInt32SmallSet(C++) with params=64, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkValidateLargeNonAscii(C++) with source=cpp-micro, suite=arrow-utf8-util-benchmarkBM_WriteColumn(C++) with params=<false,DoubleType>, source=cpp-micro, suite=parquet-arrow-reader-writer-benchmarkArrayRangeEqualsDenseUnion(C++) with params=32768/1, source=cpp-micro, suite=arrow-compare-benchmarkIsInInt16SmallSet(C++) with params=16, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkBM_RleDecoding(C++) with params=4096/1, source=cpp-micro, suite=parquet-column-io-benchmarkIsInInt64SmallSet(C++) with params=2, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkBM_ReadInt64Column(C++) with params=<Repetition::OPTIONAL, Compression::LZ4>/65536/1024, source=cpp-micro, suite=parquet-column-io-benchmarkIndexInInt32SmallSet(C++) with params=2, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkResolveManyUInt64Sorted(C++) with params=chunked_array_length:65535/num_chunks:100/num_indices:508, source=cpp-micro, suite=arrow-chunk-resolver-benchmarkIsInInt32SmallSet(C++) with params=4, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkBM_ComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkTouchArea(C++) with params=size:16777216/real_time/threads:2, source=cpp-micro, suite=arrow-memory-pool-benchmarkBM_DecodeArrowBooleanPlain(C++) with params=DecodeArrowWithNull/num_values:16384/null_in_ten_thousand:5000, source=cpp-micro, suite=parquet-encoding-benchmarkDoubleRowMajorMatrixConversionFixture(C++) with params=/ConvertToSparseCSCMatrixInt32, source=cpp-micro, suite=arrow-tensor-conversion-benchmarkListSliceInt64ListWithStepAndStop(C++) with params=32768/100, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkListSliceInt64ListWithStepAndStop(C++) with params=32768/2, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkIndexInInt32SmallSet(C++) with params=4, source=cpp-micro, suite=arrow-compute-scalar-set-lookup-benchmarkWriteFileMetadataAndData(C++) with params=num_columns:100/num_row_groups:1, source=cpp-micro, suite=parquet-metadata-benchmarkListSliceInt64ListWithStop(C++) with params=32768/2, source=cpp-micro, suite=arrow-compute-scalar-list-benchmarkBM_ReadInt64Column(C++) with params=<Repetition::REQUIRED, Compression::SNAPPY, Encoding::BYTE_STREAM_SPLIT>/1024/1024, source=cpp-micro, suite=parquet-column-io-benchmarkBM_BatchComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_HashJoinBasic_ProbeParallelism(C++) with params=Threads:8/HashTable krows:64/process_time, source=cpp-micro, suite=arrow-acero-hash-join-benchmarkWriteCsvStringRejectQuote(C++) with params=1, source=cpp-micro, suite=arrow-csv-writer-benchmarkReadCompressedBuffer(C++) with params=num_cols:1/is_partial:1/real_time, source=cpp-micro, suite=arrow-ipc-read-write-benchmarkColumnReaderSkipInt32(C++) with params=Repetition:1/BatchSize:1000, source=cpp-micro, suite=parquet-column-reader-benchmarkMemoizeLruCachedThreadUnsafe(C++) with params=64/1024, source=cpp-micro, suite=arrow-cache-benchmarkBM_ReadInt64Column(C++) with params=<Repetition::OPTIONAL, Compression::ZSTD>/1024/16, source=cpp-micro, suite=parquet-column-io-benchmarkBM_ComputeHash(C++) with params=, source=cpp-micro, suite=parquet-bloom-filter-benchmarkBM_DefinitionLevelsToBitmapRepeatedMostPresent(C++) with source=cpp-micro, suite=parquet-level-conversion-benchmark
All benchmark runs analyzed:
- Pull Request Run on
amd64-c6a-4xlarge-linuxat 2025-07-08 10:39:44Z - Pull Request Run on
arm64-t4g-2xlarge-linuxat 2025-07-08 12:09:38Z - Pull Request Run on
test-mac-armat 2025-07-08 15:00:33Z - Pull Request Run on
amd64-m5-4xlarge-linuxat 2025-07-08 11:24:41Z
Details
This report was generated using the lookback z-score method with a z-score threshold of 30.0.
Loading