Central Iowa Software Symposium
July 21 - 23, 2006 - Des Moines, IA
View the event details here ».
Chief Scientist, Relevance Inc.
Glenn Vanderburg is a principal at Relevance, where he is focused on cutting-edge software development technologies and techniques. He brings more than 20 years of experience developing software across a wide range of domains, and using a variety of tools and technologies. Glenn is always searching for ways to improve the state of software development, and was an early adopter and proponent of Ruby, Rails, and agile practices.
Performance myths about the Java platform abound, from the general "Java is slow", to the more specific "reflection is slow", "allocation is slow", "synchronization is slow", "garbage collection is slow", etc. Many of these myths have their root in fact (in JDK 1.0, everything was slow); today, not only are many of these statements not true, but Java performance has surpassed that of C in many areas, such as memory management.
In this class, we'll look at some common Java performance myths, identify where they came from, and explore the platform changes that have rendered them no longer true. Many common performance hacks don't actually help, and some can seriously hurt performance. The result is that clean code that follows common usage patterns generally shows far better behavior on modern JVMs than code laden with tweaks designed to "help" the JIT or garbage collector. More often than not, this well-intentioned assistance has the unfortunate effect of undermining many common JIT optimizations, resulting in slower -- not faster -- code.
Most of the time, Java's automatic memory management works really well—it's one of the things that makes programming in Java a pleasant and productive experience, and it's nice that we don't have to worry about managing memory manually. However, although it's usually nice to ignore memory management, occasionally we have to pay close attention. Sometimes we need to take control of certain aspects of memory management. Sometimes Java programs do exhibit memory leaks, or unacceptably long garbage collection pauses, or very poor overall performance. But because Java's memory management is supposed to be "fully automatic," it can be difficult to find out what's really going on inside the VM.
Java memory management is just like most labor-saving simplifications: it works well most of the time, but for the weird edge cases when it doesn't work quite right, it can be a nightmare. This talk opens the hood, examining the inner workings of Java's memory system, including allocation and garbage collection. We'll look at how to control the memory system and interact with it, what's costly and what's not, how to tune the garbage collector and when to switch to a different GC algorithm, and other topics.
The support infrastructure for your software project is a crucial factor for success. A new generation of tools offers significant benefits over their predecessors. This talk discusses how to choose the right mix of tools for a top-shelf project infrastructure.
The support infrastructure for your software project is a crucial factor for success. Many projects waste enormous amounts of time fighting through projects without the help of good tools. Other projects are on the right track, but could be even more successful by filling some crucial infrastructure gaps or by moving to improved tools, or by implementing policies that maximize the tools' power.
This talk looks at the latest generation of infrastructure tools, what makes them better, and how to use them well. Additionally, we'll examine the role of the infrastructure on projects and identify principles that help us understand what kinds of infrastructure we need. Tools examined will include CruiseControl, Rake, Subversion, Trac, and others.