Java Just in Time

he Java programming language promises portable, secure execution of applications. Early Java implementations relied on interpretation, leading to poor performance compared to compiled programs. Compiling Java programs to the native machine instructions provides much higher performance. Because traditional compilation would defeat Java's portability and security, another approach is necessary. This article describes some of the important issues related to just-in-time, or JIT, compilation techniques for Java. We focus on the JIT compilers developed by Sun for use with the JDK (Java Development Kit) virtual machine running on SPARC and Intel processors. (Access the Web at www.sun. com/workshop/java/jit for these compilers and additional information.) We also discuss performance improvements and limitations of JIT compilers. Future Java implementations may provide even better performance, and we outline some specific techniques that they may use. The Java Virtual Machine (JVM) assures Java application portability and security. The JVM provides a well-defined runtime framework in which Java programs are compiled for a hypothetical instruction set architecture. 1 Programs are distributed in this abstract form, divorced from the details of any other computer architecture. Running a Java program involves either interpreting JVM instructions, compiling them into instructions of the underlying hardware, or directly executing them in a hardware implementation of the JVM. The JVM is a stack machine. Each instruction gets its operands from the top of a stack, consuming those values and optionally replacing them with a result. The instructions themselves are encoded in a compact form of variable length, with the shortest instructions occupying 1 byte and most instructions being 1 to 3 bytes long. This form of encoding is known as bytecode. Previous systems such as the UCSD Pascal System and most Smalltalk implementations have used similar bytecodes. A Java source-to-bytecode compiler, such as the javac program of the JDK, compiles the classes that constitute a Java program. The compiler translates methods in each source class into bytecode instructions and places all the bytecodes for a class together in a class file. To run a Java program, the JVM loads the class file containing the program's entry point, and execution begins. The program may reference other class files, which are loaded in turn (possibly across a network). Hence, the final association of class files to form a running program takes place as execution proceeds. To maintain the integrity of the Java execution model, the JVM checks that a variety of semantic constraints are …