jorgecarleitao commented on pull request #9076: URL: https://github.com/apache/arrow/pull/9076#issuecomment-757492622
Ok, I have the benchmarks of this PR, which IMO is ready to review. FYI @alamb @nevi-me ```bash critcmp master-08cccd68802c9ddc3ca0a5d4bad6e4ba382d74b4 perf_buffer-2dca3c85920c951d92992d343b69a5384178c6fd -t 10 ``` ``` group master-08cccd68802c9ddc3ca0a5d4bad6e4ba382d74b4 perf_buffer-2dca3c85920c951d92992d343b69a5384178c6fd ----- ----------------------------------------------- ---------------------------------------------------- array_slice 512 1.17 453.4±37.97ns ? B/sec 1.00 387.0±8.80ns ? B/sec array_string_from_vec 128 1.28 4.2±0.22µs ? B/sec 1.00 3.3±0.09µs ? B/sec array_string_from_vec 256 1.45 5.9±0.14µs ? B/sec 1.00 4.1±0.12µs ? B/sec array_string_from_vec 512 1.66 9.6±0.28µs ? B/sec 1.00 5.8±0.18µs ? B/sec buffer_bit_ops and 1.23 715.9±278.78ns ? B/sec 1.00 581.4±16.32ns ? B/sec cast date64 to date32 512 1.11 7.6±0.40µs ? B/sec 1.00 6.9±0.32µs ? B/sec cast float64 to float32 512 1.21 3.2±0.10µs ? B/sec 1.00 2.7±0.06µs ? B/sec cast float64 to uint64 512 1.00 3.9±0.41µs ? B/sec 1.11 4.4±0.12µs ? B/sec cast int32 to float32 512 1.12 3.2±0.11µs ? B/sec 1.00 2.8±0.08µs ? B/sec cast int32 to int64 512 1.19 3.2±0.08µs ? B/sec 1.00 2.7±0.07µs ? B/sec cast int32 to uint32 512 1.42 3.9±0.10µs ? B/sec 1.00 2.8±0.15µs ? B/sec cast int64 to int32 512 1.96 5.1±0.27µs ? B/sec 1.00 2.6±0.07µs ? B/sec cast time32s to time64us 512 1.11 5.4±0.23µs ? B/sec 1.00 4.9±0.15µs ? B/sec concat i32 1024 1.00 4.2±0.09µs ? B/sec 1.12 4.7±0.17µs ? B/sec concat str nulls 1024 1.00 21.3±0.74µs ? B/sec 1.14 24.3±0.73µs ? B/sec eq scalar Float32 1.26 76.4±5.24µs ? B/sec 1.00 60.6±3.56µs ? B/sec filter context f32 high selectivity 1.00 309.7±1.96µs ? B/sec 1.12 346.4±10.64µs ? B/sec filter context u8 high selectivity 1.27 4.8±0.11µs ? B/sec 1.00 3.8±0.14µs ? B/sec filter context u8 w NULLs high selectivity 1.00 299.2±2.82µs ? B/sec 1.12 336.0±7.21µs ? B/sec filter u8 high selectivity 1.11 12.2±0.18µs ? B/sec 1.00 11.0±0.13µs ? B/sec from_slice prepared 1.00 834.5±26.98µs ? B/sec 1.15 963.4±36.11µs ? B/sec gt scalar Float32 1.13 42.9±1.52µs ? B/sec 1.00 37.9±1.83µs ? B/sec length 1.58 57.8±0.58µs ? B/sec 1.00 36.5±9.01µs ? B/sec like_utf8 scalar complex 1.00 1064.0±91.20µs ? B/sec 1.17 1241.6±320.92µs ? B/sec like_utf8 scalar ends with 1.14 317.0±8.58µs ? B/sec 1.00 277.5±5.80µs ? B/sec like_utf8 scalar equals 2.43 142.6±6.62µs ? B/sec 1.00 58.6±2.81µs ? B/sec like_utf8 scalar starts with 1.41 294.5±16.24µs ? B/sec 1.00 208.9±7.83µs ? B/sec min string 512 1.16 6.2±0.14µs ? B/sec 1.00 5.3±0.12µs ? B/sec mutable 1.19 587.7±15.96µs ? B/sec 1.00 492.5±17.55µs ? B/sec mutable prepared 1.17 640.6±26.75µs ? B/sec 1.00 549.0±21.62µs ? B/sec mutable str nulls 1024 1.00 4.0±0.06ms ? B/sec 1.34 5.3±0.05ms ? B/sec neq scalar Float32 1.23 76.4±3.41µs ? B/sec 1.00 61.9±6.27µs ? B/sec nlike_utf8 scalar starts with 1.00 321.1±12.95µs ? B/sec 1.27 407.8±9.19µs ? B/sec struct_array_from_vec 1024 1.53 20.8±0.55µs ? B/sec 1.00 13.6±0.66µs ? B/sec struct_array_from_vec 128 1.16 7.2±0.21µs ? B/sec 1.00 6.2±0.23µs ? B/sec struct_array_from_vec 256 1.27 9.1±0.29µs ? B/sec 1.00 7.1±0.20µs ? B/sec struct_array_from_vec 512 1.40 13.0±0.27µs ? B/sec 1.00 9.3±0.32µs ? B/sec take i32 1024 1.26 2.4±0.96µs ? B/sec 1.00 1883.9±35.60ns ? B/sec ``` ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
