alamb commented on PR #9619: URL: https://github.com/apache/arrow-rs/pull/9619#issuecomment-4565145651
The only thing that looks suspicious is some of the `list_primitive_non_null` benchmarks: ``` list_primitive_non_null/bloom_filter 1.00 397.6±39.58ms 1368.9 MB/sec 1.09 433.4±8.71ms 1255.7 MB/sec list_primitive_non_null/cdc 1.00 439.8±6.81ms 1237.3 MB/sec 1.00 438.1±7.55ms 1242.4 MB/sec list_primitive_non_null/default 1.00 261.8±3.76ms 2.0 GB/sec 1.14 297.6±8.13ms 1828.7 MB/sec list_primitive_non_null/parquet_2 1.00 293.3±1.71ms 1855.3 MB/sec 1.07 314.6±1.79ms 1730.1 MB/sec list_primitive_non_null/zstd 1.00 690.0±4.50ms 788.8 MB/sec 1.04 717.9±11.72ms 758.1 MB/sec list_primitive_non_null/zstd_parquet_2 1.00 670.3±0.95ms 812.0 MB/sec 1.05 701.4±11.29ms 776.0 MB/sec list_primitive_sparse_99pct_null/bloom_filter 1.00 11.0±0.05ms 3.3 GB/sec 1.09 12.0±0.04ms 3.1 GB/sec list_primitive_sparse_99pct_null/cdc 1.00 22.4±0.10ms 1666.3 MB/sec 1.04 23.3±0.10ms 1606.9 MB/sec list_primitive_sparse_99pct_null/default 1.00 10.6±0.04ms 3.4 GB/sec 1.09 11.6±0.04ms 3.2 GB/sec list_primitive_sparse_99pct_null/parquet_2 1.00 10.7±0.06ms 3.4 GB/sec 1.09 11.7±0.05ms 3.1 GB/sec list_primitive_sparse_99pct_null/zstd 1.00 12.5±0.05ms 2.9 GB/sec 1.07 13.4±0.04ms 2.7 GB/sec list_primitive_sparse_99pct_null/zstd_parquet_2 1.00 10.8±0.06ms 3.4 GB/sec 1.09 11.8±0.04ms 3.1 GB/sec ``` However I think @HippoBaro also saw some potential strangeness with those runs, so maybe it is a measurement thing -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
