JProfiler Get rid of your performance problems and memory leaks! Modern Garbage Collection Tuning Java Performance Training Courses COURSES AVAILABLE NOW. We can provide training courses to handle all your Java performance needs Java Performance Tuning, 2nd ed The classic and most comprehensive book on tuning Java Java Performance Tuning Newsletter Your source of Java performance news. Subscribe now! Enter email: Training online Threading Essentials course JProfiler Get rid of your performance problems and memory leaks!

Back to newsletter 144 contents

A few months ago I discussed initial and maximum heaps sizes, feeding back to you the aggregate responses I got from you on the subject. Interestingly, in one of this month's referenced articles about tuning WebSphere Application Server (see the articles below), IBM's own performance experts state clearly that for IBM JVMs the initial and max values should be the same for optimal performance - which is also what I said in that previous newsletter based on your feedback. Bear in mind that this advice is specific to the IBM JVM, and should not be generalised to other JVMs - each is different, and even within one JVM the garbage collector used can affect the validity of this choice. The Oracle JVM did not appear to show optimal performance when initial equals max.

A note from this newsletter's sponsor

The example IBM application tested showed a huge difference in performance from setting initial to max - from 8% of the time being spent in GC down to 2%. That's extraordinary. Though I always temper my response to large gains like this by remembering what my first boss said to me when I reported a 50-fold improvement in performance after a change: "wow, I didn't realise the peformance was THAT bad before!".

Now I'm going to do something I almost never do: mention a commercial tool here in the editorial section for no reason other than that it looks particularly interesting. But it is free even when used in prod, so I think it's reasonable. The tool? Nastel's new AutoPilot Heap Detective. A free tool that helps you analyse your JVM memory with no overhead except when (infrequently) used, sounds almost too good to be true - unexpected memory issues in production is one of the more painful things to happen, and the easier, quicker and earlier you can diagnose these types of problems, the better. Have a look at the tools section below if you want more details on their tool.

Now on to all our usual links to Java performance tools, news, articles and, as ever, all the extracted tips from all of this month's referenced articles.

A note from this newsletter's sponsor

News

Java performance tuning related news.

• Real World Java EE Patterns - Rethinking Best Practices
• Nastel announces their free Java memory leak detector, AutoPilot Heap Detective
• Announcing Batoo JPA, built on top of JPA 2.0 spec with performance being the top priority
• Five Myths about Hash Tables
• Top 5 Virtualization Performance Myths

Tools

Java performance tuning related tools.

• AutoPilot Heap Detective production ready memory analyzer
• batoo "The Fastest JPA Implementation"
• JAMon (Java Application Monitor)
• Java Simon - Simple Monitoring API
• Highly Scalable Java: A collection of Concurrent and Highly Scalable Utilities

A note from this newsletter's sponsor

Articles

• OutOfMemoryError: unable to create new native thread - Problem Demystified (Pierre-Hugues Charbonneau/Java EE Support Patterns)
• 10 Garbage Collection Interview Questions and Answers (Javin Paul/javarevisited)
• Case study: Tuning WebSphere Application Server V7 and V8 for performance (David Hare, Christopher Blythe/IBM)
• Five ways to maximize Java NIO and NIO.2 (Cameron Laird/JavaWorld)
• Compact Off-Heap Structures/Tuples In Java (Martin Thompson/Mechanical Sympathy)
• Back to Basics: The Theory of (Performance) Relativity (Lori MacVittie/DevCentral)

Jack Shirazi

Back to newsletter 144 contents