[ https://issues.apache.org/jira/browse/HBASE-12213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14612998#comment-14612998 ]
Anoop Sam John commented on HBASE-12213: ---------------------------------------- bq.We get around 9 to 10 % improvement. That is great Ram.. Thanks for doing the perf test for every sub task.. Yes the MBB and usage of its we tried to save every op as much as possible.. When we avoid the temp on heap buffer creation and copy and MBB actually wraps the underlying offheap BBs only, we will be able to see much more gain. Random performance is the main thing we are expecting there too. > HFileBlock backed by Array of ByteBuffers > ----------------------------------------- > > Key: HBASE-12213 > URL: https://issues.apache.org/jira/browse/HBASE-12213 > Project: HBase > Issue Type: Sub-task > Components: regionserver, Scanners > Reporter: Anoop Sam John > Assignee: ramkrishna.s.vasudevan > Attachments: HBASE-12213_1.patch, HBASE-12213_2.patch > > > In L2 cache (offheap) an HFile block might have been cached into multiple > chunks of buffers. If HFileBlock need single BB, we will end up in recreation > of bigger BB and copying. Instead we can make HFileBlock to serve data from > an array of BBs. -- This message was sent by Atlassian JIRA (v6.3.4#6332)