tyrelr commented on pull request #9215: URL: https://github.com/apache/arrow/pull/9215#issuecomment-766361051
I'm still seeing a mix of inconsistent performance hits/bumps after the rebase. ``` critcmp master-67d0c2e38 array-slice-83b8938af -t 10 group array-slice-83b8938af master-67d0c2e38 ----- --------------------- ---------------- array_slice 512 1.11 127.8±0.36ns ? B/sec 1.00 115.4±0.15ns ? B/sec cast int64 to int32 512 1.00 2.3±0.00µs ? B/sec 1.11 2.5±0.01µs ? B/sec concat i32 1024 1.00 2.5±0.00µs ? B/sec 1.19 3.0±0.01µs ? B/sec equal_512 1.14 46.7±0.18ns ? B/sec 1.00 41.1±0.05ns ? B/sec like_utf8 scalar complex 1.00 1075.9±2.54µs ? B/sec 1.16 1249.6±2.29µs ? B/sec like_utf8 scalar equals 1.00 70.8±0.08µs ? B/sec 1.25 88.2±0.12µs ? B/sec min nulls string 512 1.14 6.5±0.04µs ? B/sec 1.00 5.7±0.03µs ? B/sec min string 512 1.00 3.4±0.00µs ? B/sec 1.28 4.3±0.01µs ? B/sec multiply 512 1.38 346.6±0.37ns ? B/sec 1.00 251.2±0.49ns ? B/sec nlike_utf8 scalar complex 1.00 1162.1±1.30µs ? B/sec 1.12 1300.7±1.45µs ? B/sec take bool nulls 1024 1.00 3.7±0.03µs ? B/sec 1.38 5.1±0.03µs ? B/sec take bool nulls 512 1.00 1720.3±8.69ns ? B/sec 1.42 2.5±0.02µs ? B/sec take i32 512 1.11 1023.5±1.61ns ? B/sec 1.00 918.6±1.07ns ? B/sec take i32 nulls 512 1.00 989.6±1.31ns ? B/sec 1.10 1089.6±2.06ns ? B/sec ``` I won't push my rebase up unless we decide on some further tweaks to make, as it doesn't seem worth forcing a re-review since there were no conflicts. ---------------------------------------------------------------- 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]
