[
https://issues.apache.org/jira/browse/IGNITE-3640?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15686963#comment-15686963
]
Taras Ledkov edited comment on IGNITE-3640 at 11/22/16 3:01 PM:
----------------------------------------------------------------
Here are results with our simple IgfsBenchmark:
Buffered read: *7048 +/- 125*
Not buffered read: *7150 +/- 127*
Look like the difference is less than dispersion. But the benchmark checks only
serial read file from begin to EOF.
May be if we check repeatable read with changing position inside the buffer the
difference will be more significant.
Should we check it?
was (Author: tledkov-gridgain):
Here are results with our simple IgfsBenchmark:
Buffered read: *7048 +/- 125*
Not buffered read: *7150 +/- 127*
Look like the difference is less then dispersion. But the benchmark checks only
serial read file from begin to EOF.
May be if we check repeatable read with changing position inside the buffer the
difference will be more significant.
Should we check it?
> IGFS: Local secondary: Investigate whether we need BufferedInputStream for
> open operation.
> ------------------------------------------------------------------------------------------
>
> Key: IGNITE-3640
> URL: https://issues.apache.org/jira/browse/IGNITE-3640
> Project: Ignite
> Issue Type: Task
> Components: IGFS
> Affects Versions: 1.6
> Reporter: Vladimir Ozerov
> Assignee: Taras Ledkov
> Fix For: 2.0
>
>
> We need to accurately investigate IGFS input stream workflow. I have doubts
> we really need to extend from {{BufferedInputStream}} here.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
