spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Dongjoon Hyun (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SPARK-16699) Fix performance bug in hash aggregate on long string keys
Date Mon, 25 Jul 2016 01:37:20 GMT

    [ https://issues.apache.org/jira/browse/SPARK-16699?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15391239#comment-15391239
] 

Dongjoon Hyun commented on SPARK-16699:
---------------------------------------

Hi, [~qifan].
Nice catch! By the way, usually, only committers set `FIX VERSION` field.
You had better leave it blank next time. :)

> Fix performance bug in hash aggregate on long string keys
> ---------------------------------------------------------
>
>                 Key: SPARK-16699
>                 URL: https://issues.apache.org/jira/browse/SPARK-16699
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core
>    Affects Versions: 2.0.0
>            Reporter: Qifan Pu
>             Fix For: 2.0.0
>
>
> In the following code in `VectorizedHashMapGenerator.scala`:
> ```
>     def hashBytes(b: String): String = {
>       val hash = ctx.freshName("hash")
>       s"""
>          |int $result = 0;
>          |for (int i = 0; i < $b.length; i++) {
>          |  ${genComputeHash(ctx, s"$b[i]", ByteType, hash)}
>          |  $result = ($result ^ (0x9e3779b9)) + $hash + ($result << 6) + ($result
>>> 2);
>          |}
>        """.stripMargin
>     }
> ```
> when b=input.getBytes(), the current 2.0 code results in getBytes() being called n times,
n being length of input. getBytes() involves memory copy is thus expensive and causes a performance
degradation.
> Fix is to evaluate getBytes() before the for loop.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org


Mime
View raw message