You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@kylin.apache.org by "liyang (JIRA)" <ji...@apache.org> on 2015/11/25 08:05:11 UTC

[jira] [Created] (KYLIN-1167) Vector processing in coprocessor

liyang created KYLIN-1167:
-----------------------------

             Summary: Vector processing in coprocessor
                 Key: KYLIN-1167
                 URL: https://issues.apache.org/jira/browse/KYLIN-1167
             Project: Kylin
          Issue Type: Bug
            Reporter: liyang
            Assignee: Dong Li


Some use has experience slow HBase processing when the scan range is over 20+ million records.

To improve coprocessor, one way is to process records in vector (instead of one by one). Previous works by Power Drill (Google) and Tungsten (Spark) all mentioned this technique. Said it can improve processing speed by 4 times or so.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)