[
https://issues.apache.org/jira/browse/HBASE-15536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15221395#comment-15221395
]
Duo Zhang edited comment on HBASE-15536 at 4/1/16 8:50 AM:
-----------------------------------------------------------
So we are using more CPUs with asyncwal? But actually we spend less cycles with
asyncwal(Yeah you said, less work)? What does the '2.597 GHz' mean?
And for the IPC, I remember the time when I was a graduate student who was
learning how to do 'software pipeline' and 'instruction scheduling' in a
compiler... But usually these optimizations are only useful on loops and I can
not find a heavy loop in {{AsyncFSWAL}} (maybe the append loop in consume
method?)... And it is interesting that we have a less branch-miss rate but a
less IPC...
Maybe we need to find the hot spots in {{AsyncFSWAL}} first?
Thanks.
was (Author: apache9):
So we are using more CPUs with asyncwal? But actually we spend less cycles with
asyncwal(Yeah you said, less work)? What does the '2.597 GHz' mean?
And for the IPC, I remember the time when I was a graduate student who was
learning how to do 'software pipeline' and 'instruction scheduling' in a
compiler... But usually these optimizations are only useful on loops and I can
not find a heavy loop in {{AsyncFSWAL}}... And it is interesting that we have a
less branch-miss rate but a less IPC...
Maybe we need to find the hot spots in {{AsyncFSWAL}} first?
Thanks.
> Make AsyncFSWAL as our default WAL
> ----------------------------------
>
> Key: HBASE-15536
> URL: https://issues.apache.org/jira/browse/HBASE-15536
> Project: HBase
> Issue Type: Sub-task
> Reporter: Duo Zhang
>
> As it should be predicated on passing basic cluster ITBLL
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
