You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@commons.apache.org by "Amir Hadadi (JIRA)" <ji...@apache.org> on 2019/03/04 21:06:00 UTC
[jira] [Updated] (TEXT-153) LookupTranslator performance
optimization
[ https://issues.apache.org/jira/browse/TEXT-153?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amir Hadadi updated TEXT-153:
-----------------------------
Description:
When using Java Mission Control to profile an application that uses StringEscapeUtils::escapeEcmaScript, I noticed that a lot of time is spent in LookupTranslator::translate at the prefixSet::contains check.
I suggest taking advantage of the fact that prefixSet contains only characters, and replace it with a BitSet.
I did some benchmarking and translate for the non escaped case is 4-5X faster when replacing the HashSet with a BitSet.
BitSet Memory consumption for characters is capped at 8KB, and depends on the maximal prefix character. For example for ECMA script the max escaped prefix character is "\" which has unicode code 92 so the BitSet uses a long array of length 2 to represent all the needed characters.
Link to pull request: https://github.com/apache/commons-text/pull/108
was:
When using Java Mission Control to profile an application that uses StringEscapeUtils::escapeEcmaScript, I noticed that a lot of time is spent in LookupTranslator::translate at the prefixSet::contains check.
I suggest taking advantage of the fact that prefixSet contains only characters, and replace it with a BitSet.
I did some benchmarking and translate for the non escaped case is 4-5X faster when replacing the HashSet with a BitSet.
BitSet Memory consumption for characters is capped at 8KB, and depends on the maximal prefix character. For example for ECMA script the max escaped prefix character is "\" which has unicode code 92 so the BitSet uses a long array of length 2 to represent all the needed characters.
> LookupTranslator performance optimization
> -----------------------------------------
>
> Key: TEXT-153
> URL: https://issues.apache.org/jira/browse/TEXT-153
> Project: Commons Text
> Issue Type: Improvement
> Affects Versions: 1.0
> Reporter: Amir Hadadi
> Priority: Minor
> Time Spent: 10m
> Remaining Estimate: 0h
>
> When using Java Mission Control to profile an application that uses StringEscapeUtils::escapeEcmaScript, I noticed that a lot of time is spent in LookupTranslator::translate at the prefixSet::contains check.
> I suggest taking advantage of the fact that prefixSet contains only characters, and replace it with a BitSet.
> I did some benchmarking and translate for the non escaped case is 4-5X faster when replacing the HashSet with a BitSet.
> BitSet Memory consumption for characters is capped at 8KB, and depends on the maximal prefix character. For example for ECMA script the max escaped prefix character is "\" which has unicode code 92 so the BitSet uses a long array of length 2 to represent all the needed characters.
>
> Link to pull request: https://github.com/apache/commons-text/pull/108
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)