[
http://jira.jboss.com/jira/browse/JBCACHE-937?page=comments#action_12357472 ]
Manik Surtani commented on JBCACHE-937:
---------------------------------------
This shows the VAM to be (much) faster than just serializing the method call, although the
VAM + Java Serialization is quicker than the VAM + JBoss Serialization. I'm waiting
on some comments from Clebert about this, although here is his initial response:
"in recent Java5 (they have ported the fix to java4 as well) they have
fixed the synchronized caching what was one of the advantages of jbser.
(don't know if pressured by jboss-serialization or not)."
We need to decide whether it still makes sense to ship with JBoss Serialization, or to
simply recommend using a JVM with synchronized caching.
Benchmark VersionAwareMarshaller
--------------------------------
Key: JBCACHE-937
URL:
http://jira.jboss.com/jira/browse/JBCACHE-937
Project: JBoss Cache
Issue Type: Task
Security Level: Public(Everyone can see)
Affects Versions: 2.0.0.ALPHA2
Reporter: Galder Zamarreno
Assigned To: Manik Surtani
Fix For: 2.0.0.GA, 2.0.0.CR1
Benchmark VersionAwareMarshaller to make it more performant. There's already a
unit test to use it as base: VersionAwareMarshallerPerfTest.
It should probably make comparison failures based on results from different
marshalling methods, i.e, if difference bigger than x, fail. Also add marshalling
different
types of objects.
Potential performance improvements:
- Using similar techniques as used in ExposedByteArrayOutputStream (JGroups) for VAM?
Needs profiling to determine effect.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira