[
https://issues.apache.org/jira/browse/HIVE-9153?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14251597#comment-14251597
]
Rui Li commented on HIVE-9153:
------------------------------
Judging from the results, I think fewer mappers can improve overall
performance, which is true for both spark and tez. Problem is that, why spark
is 60s slower than tez with same # of mappers.
One possible reason is that we don't have data locality with
CombineHiveInputFormat, which is tracked by HIVE-8722.
I also noticed that the parallelism drops during execution (attach a screenshot
later). This may be due to the delay schedule mechanism of spark, which
attempts to schedule tasks with some locality first.
> Evaluate CombineHiveInputFormat versus HiveInputFormat [Spark Branch]
> ---------------------------------------------------------------------
>
> Key: HIVE-9153
> URL: https://issues.apache.org/jira/browse/HIVE-9153
> Project: Hive
> Issue Type: Sub-task
> Components: Spark
> Affects Versions: spark-branch
> Reporter: Brock Noland
> Assignee: Rui Li
> Attachments: screenshot.PNG
>
>
> The default InputFormat is {{CombineHiveInputFormat}} and thus HOS uses this.
> However, Tez uses {{HiveInputFormat}}. Since tasks are relatively cheap in
> Spark, it might make sense for us to use {{HiveInputFormat}} as well. We
> should evaluate this on a query which has many input splits such as {{select
> count(\*) from store_sales where something is not null}}.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
