[MCOL-1409] Optimisation of mcsapi Created: 2018-05-13  Updated: 2022-11-05  Resolved: 2022-11-05

Status: Closed
Project: MariaDB ColumnStore
Component/s: N/A
Affects Version/s: None
Fix Version/s: Icebox

Type: Task Priority: Major
Reporter: Andrew Hutchings (Inactive) Assignee: Unassigned
Resolution: Won't Do Votes: 1
Labels: None

Epic/Theme: MCOL-4571

 Description   

Largest amount of CPU time appears to be single byte copies in weBulkInsert() to messageIn. If we flip the array of table->rows we can push an entire column of rows in one go rather than an entry at a time. This should accelerate the writeRow() command.



 Comments   
Comment by Andrew Hutchings (Inactive) [ 2018-06-25 ]

Better fix for this:

Engine: modify colValueList (and dictStrList) so that the vector is [row][column]
API: setInsert inserts directly into the network buffer instead of arrays that need converting

Comment by Todd Stoffel (Inactive) [ 2022-11-05 ]

Item is out of date. Closing due to inactivity. If you feel this was done in error please open a new ticket.

Generated at Thu Feb 08 02:28:33 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.