[SERVER-61843] Figure out what is going on with our insert_vector performance Created: 01/Dec/21  Updated: 27/Oct/23  Resolved: 04/Jan/22

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Matthew Russotto Assignee: Backlog - Replication Team
Resolution: Gone away Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
related to SERVER-62193 Improve insert_vector secondary perfo... Open
Assigned Teams:
Replication
Operating System: ALL
Participants:

 Description   

In investigating SERVER-61792 I happened to notice that performance on our insert_vector benchmarks has gone down significantly since v4.4. I took a quick look and traced some of that to SERVER-39950, but it might be worth seeing if we can recover most or all of it.



 Comments   
Comment by Matthew Russotto [ 21/Dec/21 ]

SERVER-58949 causes a few percent decrease in performance, probably by doubling the recordid size. This could possibly be avoided by aliasing the SharedBuffer (12 bytes, probably) with the char buffer in RecordId, but I'm not sure it's worth it.

Generated at Thu Feb 08 05:53:27 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.