Jvm performance tuning (notes)

I'm the author of this post. Note that it's entirely adapted from the talk done by Attila Szegedi (<a href="http://www.infoq.com/presentations/JVM-Performance-Tuning-twitter" rel="nofollow">http://www.infoq.com/presentations/JVM-Performance-Tuning-tw...</a>), who deserves all the credit. I just filled in some of the gaps and reworked it so it flows in textual form.<p>I'll try to update the post with any comments I get here. It's definitely not ground truth as I was a bit fuzzy on some parts of the presentation.

Great article. Note that the point about Strings using 2 bytes per char isn't always true. The latest HotSpot VM has an option called "-XX:+UseCompressedStrings" (see <a href="http://www.oracle.com/technetwork/java/javase/tech/vmoptions-jsp-140102.html" rel="nofollow">http://www.oracle.com/technetwork/java/javase/tech/vmoptions...</a>) where the VM will try to use ASCII-strings whenever possible to save memory.

&#62; Don’t write your own memory manager, and stop if you find yourself doing something ugly with byte buffers.<p>What's wrong with byte buffers? Serious question, as I use them all the time whenever I do OpenGL-ES 2.0 surfaceviews for android apps.

FTA: <i>Twitter had a service that would have a terrible GC pause every three days. Solution: just bounce the machine after less than three days.</i><p>That is not what I would call a solution. It's a <i>work-around</i>.<p>And not a very good one.