Lars Bak received the MS degree in computer science from Aarhus University in 1988. As one of the founders of Mjølnevr Informatics, he participated in the design and implementation of the Beta Mjølner System. In 1991, he joined the Self group at Sun Microsystems Laboratories. During his time there, he developed a programming environment for Self and added several enhancements to the virtual machine. In 1994, he joined Longiew Technologies LLC, where he designed and implemented high performance virtual machines for both Smalltalk and Java. Longview was acquired by Sun Microsystems, Inc. in 1997. Lars is presently engineering manager and technical lead in the HotSpot team at Java Software Division, Sun Microsystems, Inc.
The Java HotSpot™ Performance Engine is a high performance virtual machine for Java. It achieves fast startup, small footprint, and high performance by combining on-the-fly compilation, fast thread synchronization, and generational garbage collection. This presentation will include a discussion of the key technologies of HotSpot and conclude with a demonstration.
A major attraction of the Java programming language is that it provides automatic memory management (garbage collection). This frees the programmer from inserting explicit memory deallocation statements into programs and it helps guarantee “safe” execution semantics. Traditionally, garbage collection has been considered an inefficient process that slowed down program execution and introduced disruptive pauses. The Java HotSpot™ Performance Engine solves these problems by employing state-of-the-art garbage collection techniques ensuring fast allocation and non-disruptive pauses. This results in improved throughput and responsiveness for enterprise server applications and pause-less interactive applications. The presentation will, in details, describe the architecture and the benefits of the HotSpot garbage collector.
HotSpot Past, Present, and Future Performance (30 min) Performance for a Java™ virtual machine can be categorized into four areas: startup time, peak performance, interactive performance, and memory footprint. These categories will be used to describing the evolution of both the Java HotSpot™ Performance Engine and the Java HotSpot™ Client Virtual Machine. In addition, future performance improvement will be discussed.