Actions
Support #1915
closed![Author: Philip Fearon](https://www.gravatar.com/avatar/dd3c76e6c50a14203d6d7d6b989425fc?rating=PG&size=50&default=https%3A%2F%2Fassets.plan.io%2Fimages%2Fdefault_avatar.png)
![Assignee: Debbie Lockett](https://www.gravatar.com/avatar/9c2ec304a5bec55af709863d87d5f6b7?rating=PG&size=22&default=https%3A%2F%2Fassets.plan.io%2Fimages%2Fdefault_avatar.png)
Test case for Saxon XSLT Performance comparison
Start date:
2013-10-01
Due date:
% Done:
0%
Estimated time:
Legacy ID:
Applies to branch:
Fix Committed on Branch:
Fixed in Maintenance Release:
Platforms:
Description
This test case for XSLT shows a significant difference in performance for different Saxon versions
so I thought it might be of interest. On Saxon 9.5.1.2 the test runs 90 time faster than for Saxon 9.3.1.1.
I just ran the XSLT from the command line (in Windows 8) as follows:
java -cp "J:\SaxonPE9-4-0-9J\saxon9pe.jar" net.sf.saxon.Transform -t -xsl:xquery2xml.xsl -s:input-file1.xml -o:output/result.xml
Performance results are summarised below:
Version: Time (ms)
SaxonPE9-3-0-11J 1899939
SaxonPE9-4-0-9J 2857
SaxonPE9-5-1-2J 2106
This is a non-typical test case in that the XSLT stylesheet functions as a lexer for XQuery code.
The input is initially processed as a sequence of unicode codepoints, but a lengthy regular
expression is used later to tokenize text that is not within XQuery comments or string-literals.
There's also a lot of tail-recursion going on.
Files
Please register to edit this issue
Actions