Characteristics of dynamic JVM languages

Abstract

The Java Virtual Machine (JVM) has become an execution platform targeted by many programming languages. However, unlike with Java, a statically-typed language, the performance of the JVM and its Just-In-Time (JIT) compiler with dynamically-typed languages lags behind purpose-built language-specific JIT compilers. In this paper, we aim to contribute to the understanding of the workloads imposed on the JVM by dynamic languages. We use various metrics to characterize the dynamic behavior of a variety of programs written in three dynamic languages (Clojure, Python, and Ruby) executing on the JVM. We identify the differences with respect to Java, and briefly discuss their implications.

Workload characterization of JVM languages

“…We note that (based on our results) we cannot evaluate whether any actual optimizations would pay off-optimizing a program involves many trade-offs-and not all opportunities can be exploited. Nevertheless, we believe that the results of our investigation are worth the developer's attention and that highlighting characteristic differences in workloads can provide directions for further study.In general, this article consolidates our previous research on workload characterization [4][5][6][7][8]. To add an additional perspective to our study, we also present an analysis of the ability of the HotSpot JVM to perform a fundamental optimization, inlining, on the workloads written in different JVM languages.…”

“…In general, this article consolidates our previous research on workload characterization . To add an additional perspective to our study, we also present an analysis of the ability of the HotSpot JVM to perform a fundamental optimization, inlining, on the workloads written in different JVM languages.…”

Cross-language compiler benchmarking: are we fast yet?

“…We use 10 language implementations of which four are implemented on the Graal and Truffle platform. These languages are JavaScript with Oracle's Graal.js 11 implementation, Ruby with JRuby+Truffle, 12 Newspeak with SOMns, as well as Smalltalk with Truffle-SOM. 13 To complement these languages, we include Java 8 based on Oracle's HotSpot JVM as a baseline.…”

“…Other work focuses on characterizing differences of languages. Sarimbekov et al [12] examine the CLBG benchmarks to determine their behavior for Clojure, Java, JRuby, and Jython based on dynamic metrics. Li et al [8] do a similar study adding Scala and additional applications for each language.…”

Approaches to interpreter composition

“…While Java programs on HotSpot have excellent performance, other languages (e.g. Python) on HotSpot often run slower than simple C-based interpreters [8,39]. While better VM extensions (e.g.…”