[YSQL] [Parallel Query] Few Orderby microbenchmarks regressed by 35-58% with the feature enabled

[shantanugupta-yb]

Jira Link: DB-10559

Description

Description Comparison is between:

• Orderby workload executed without parallel workers and without analyse executed on tables
• Orderby workload executed without parallel workers and with analyse executed on tables
• Orderby workload executed with parallel workers and with analyse executed on tables

Report: https://perf.dev.yugabyte.com/report/view/W3sibmFtZSI6Ik9yZGVyYnkgZGVmYXVsdCArIGFuYWx5c2UiLCJ0ZXN0X2lkIjoiNTAyNzUwMiIsImlzQmFzZWxpbmUiOnRydWV9LHsibmFtZSI6Ik9yZGVyYnkgcGFyYWxsZWwgKyBhbmFseXNlIiwidGVzdF9pZCI6IjUwMjc2MDIiLCJpc0Jhc2VsaW5lIjpmYWxzZX0seyJuYW1lIjoiT3JkZXJieSBkZWZhdWx0IiwidGVzdF9pZCI6IjQ5ODYyMDIiLCJpc0Jhc2VsaW5lIjpmYWxzZX1d#report-table

Orderby microbenchmarks which have regressed beyond 35-58%

Issue Type

kind/bug

Warning: Please confirm that this issue does not contain any sensitive information

• [X] I confirm this issue does not contain any sensitive information.

[rthallamko3]

@shantanugupta-yb , Can you get the baseline with CBO gflag + analyze and post the comparison?

[rthallamko3]

Orderby: Report

[rthallamko3]

PG did not use parallelism, but YB used parallelism. YB with parallelism is slightly slower than without parallelism.

cc @andrei-mart

[rthallamko3]

cc @tverona1 , @gauravk-in as this has impact on cost modeling discussions with Parallel Query. Could be related to https://github.com/yugabyte/yugabyte-db/issues/21670