Initialize once, start fast: application initialization at build time

Abstract

Arbitrary program extension at run time in language-based VMs, e.g., Java's dynamic class loading, comes at a startup cost: high memory footprint and slow warmup. Cloud computing amplifies the startup overhead. Microservices and serverless cloud functions lead to small, self-contained applications that are started often. Slow startup and high memory footprint directly affect the cloud hosting costs, and slow startup can also break service-level agreements. Many applications are limited to a prescribed set of pre-tested classes, i.e., use a closed-world assumption at deployment time. For such Java applications, GraalVM Native Image offers fast startup and stable performance. GraalVM Native Image uses a novel iterative application of points-to analysis and heap snapshotting, followed by ahead-of-time compilation with an optimizing compiler. Initialization code can run at build time, i.e., executables can be tailored to a particular application configuration. Execution at run time starts with a pre-populated heap, leveraging copy-on-write memory sharing. We show that this approach improves the startup performance by up to two orders of magnitude compared to the Java HotSpot VM, while preserving peak performance. This allows Java applications to have a better startup performance than Go applications and the V8 JavaScript VM. CCS Concepts: • Software and its engineering → Runtime environments.

Scalable pointer analysis of data structures using semantic models

“…Micronaut, Quarkus, and Helidon are all frameworks for building cloud based applications in Java. They were chosen as Java is extensively used in this domain as well as because the past work [55] on GraalVM Native Image uses them for evaluation. The remaining benchmarks are general-purpose applications spanning multiple diverse domains chosen from the set of applications the original proposal for semantic models [15] evaluates on.…”

“…GraalVM Native Image [34,35,55] is an ahead-of-time compiler for Java applications built on top of the GraalVM compiler [11,12,14,27]. It optimizes for start up time and memory footprint of the generated binaries.…”

“…GraalVM Native Image makes a closed world assumption to enable ahead of time compilation, thus requiring the user to list, at build time, all uses of language features like reflection, native code access, and unsafe memory accesses (via sun.misc.Unsafe in Java). The interested reader is referred to [55] for further details regarding GraalVM Native Image.…”

Execution Migration of Managed Languages Programs Among Heterogeneous-ISA Processing Units

“…Because GraalVM Native Image runs Java applications faster or as fast as Java HotSpot VM [37], and Popcorn adds near to zero overhead to applications [4], our design is likely to be of maximum performance.…”

“…From Java to LLVM IR GraalVM is a high-performance language runtime for JVM-based languages, JavaScript, Ruby, Python, and a number of other popular languages. Its subproject, GraalVM Native Image [37], allows AOT compilation of Java code to a standalone executable binary (native image, or machine code). This executable includes the application classes, dependencies, runtime library, and other machine code from the JDK.…”