[
https://issues.jboss.org/browse/JBRULES-2764?page=com.atlassian.jira.plug...
]
Tihomir Surdilovic commented on JBRULES-2764:
---------------------------------------------
We focused on improving the session creation time and test show following data compared to
provided 511 tests (all test ran locally):
run501.sh
RuleTest: count=10000, total=1677ms, average=0.1677ms
KnowledgeTest: count=10000, total=1385ms, average=0.1385ms
***********
run511.sh
RuleTest: count=10000, total=4211ms, average=0.4211ms
KnowledgeTest: count=10000, total=3820ms, average=0.382ms
***********
run520-snapshot.sh
RuleTest: count=10000, total=105ms, average=0.0105ms
KnowledgeTest: count=10000, total=86ms, average=0.0086ms
and exec tests:
exec501.sh
RuleTest: count=10000, total=3418ms, average=0.3418ms
KnowledgeTest: count=10000, total=2941ms, average=0.2941ms
***********
exec511.sh
RuleTest: count=10000, total=6000ms, average=0.6ms
KnowledgeTest: count=10000, total=5976ms, average=0.5976ms
***********
exec520snapshot.sh
RuleTest: count=10000, total=1865ms, average=0.1865ms
KnowledgeTest: count=10000, total=4533ms, average=0.4533ms
Test show significant improvements on the session creation time. With session execution we
are limited as to what can be done as there were significant new features added to Drools
5.1. Not sure what can be done about that short-term.
Let me know if these tests results would suite the needs.
Thanks.
Performance degradation from 4.0.7 to 5.0.1 to 5.1.1
----------------------------------------------------
Key: JBRULES-2764
URL:
https://issues.jboss.org/browse/JBRULES-2764
Project: Drools
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: drools-core, drools-core (expert)
Affects Versions: 5.0.1.FINAL, 5.1.1.FINAL
Environment: Sun JDK 1.6.0_20 on Fedora 11 (Linux
2.6.30.10-105.2.23.fc11.i686.PAE)
Reporter: David Ward
Assignee: Mark Proctor
I have noticed a performance degradation from 4.0.7 to 5.0.1, and 5.0.1 to 5.1.1. Both
in creating new StatelessSessions, and executing them.
I am attaching a portable test case which shows the degradation. Please refer to the
test.zip!/README.txt file for full instructions and explanation.
For the purpose of this jira description (and to show the degradation from a high-level),
here is sample output. Note that the run*.sh scripts only create new StatelessSessions,
but do not execute them. The exec*.sh scripts also execute the StatelessSessions.
run407.sh
RuleTest: count=10000, total=33ms, average=0.0033ms
KnowledgeTest: API not available
run501.sh
RuleTest: count=10000, total=1147ms, average=0.1147ms
KnowledgeTest: count=10000, total=784ms, average=0.0784ms
run511.sh
RuleTest: count=10000, total=2381ms, average=0.2381ms
KnowledgeTest: count=10000, total=2007ms, average=0.2007ms
exec407.sh
RuleTest: count=10000, total=875ms, average=0.0875ms
KnowledgeTest: API not available
exec501.sh
RuleTest: count=10000, total=2641ms, average=0.2641ms
KnowledgeTest: count=10000, total=2223ms, average=0.2223ms
exec511.sh
RuleTest: count=10000, total=4188ms, average=0.4188ms
KnowledgeTest: count=10000, total=3664ms, average=0.3664ms
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira