hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lars Hofhansl (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-13448) New Cell implementation with cached component offsets/lengths
Date Wed, 20 May 2015 00:34:00 GMT

    [ https://issues.apache.org/jira/browse/HBASE-13448?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14551555#comment-14551555

Lars Hofhansl commented on HBASE-13448:

My 0.98 patch must be wrong.
Am seeing it taking more time with the patch than without (I'm doing runs of 3 and measure
the standard deviation, so it's not a test anomaly at least).

10m rows, 5 cols each, 1 version, 8 byte keys, 8 byte values.
Scan scanning over all rows filtering everything at the server with a ValueFilter. Numbers
in seconds (mean and standard deviation):

||tracker||w/ patch||w/o patch||
|explicit|m=7.5 s=0.01|m=7.03 s=0.006|
|wild|m=6.278 s=0.005|m=5.59 s=0.004|

Not trusting my 0.98 patch now. Will take another look at it. And I will also test with only
caching the row length.

> New Cell implementation with cached component offsets/lengths
> -------------------------------------------------------------
>                 Key: HBASE-13448
>                 URL: https://issues.apache.org/jira/browse/HBASE-13448
>             Project: HBase
>          Issue Type: Sub-task
>          Components: Scanners
>            Reporter: Anoop Sam John
>            Assignee: Anoop Sam John
>             Fix For: 2.0.0
>         Attachments: 13291-0.98.txt, HBASE-13448.patch, HBASE-13448_V2.patch, HBASE-13448_V3.patch,
gc.png, hits.png
> This can be extension to KeyValue and can be instantiated and used in read path.

This message was sent by Atlassian JIRA

View raw message