Java Virtual Machine Research Articles

Sharing Programming Resources Between Bio* Projects.

Open-source software encourages computer programmers to reuse software components written by others. In evolutionary bioinformatics, open-source software comes in a broad range of programming languages, including C/C++, Perl, Python, Ruby, Java, and R. To avoid writing the same functionality multiple times for different languages, it is possible to share components by bridging computer languages and Bio* projects, such as BioPerl, Biopython, BioRuby, BioJava, and R/Bioconductor.In this chapter, we compare the three principal approaches for sharing software between different programming languages: by remote procedure call (RPC), by sharing a local “call stack,” and by calling program to programs. RPC provides a language-independent protocol over a network interface; examples are SOAP and Rserve. The local call stack provides a between-language mapping, not over the network interface but directly in computer memory; examples are R bindings, RPy, and languages sharing the Java virtual machine stack. This functionality provides strategies for sharing of software between Bio* projects, which can be exploited more often.Here, we present cross-language examples for sequence translation and measure throughput of the different options. We compare calling into R through native R, RSOAP, Rserve, and RPy interfaces, with the performance of native BioPerl, Biopython, BioJava, and BioRuby implementations and with call stack bindings to BioJava and the European Molecular Biology Open Software Suite (EMBOSS).In general, call stack approaches outperform native Bio* implementations, and these, in turn, outperform “RPC”-based approaches. To test and compare strategies, we provide a downloadable Docker container with all examples, tools, and libraries included.

Towards efficient XML parsing through minimization of JVM parameter space

A significant increase in the usage of Extensible Markup Language (XML) data for various protocols and standards emphasizes the development of efficient XML parsers. For the Java language, the XML DOM parser despite performing in-memory operations is unable to achieve peak execution performance on modern systems, especially for parsing large XML files. The issue of inefficient execution may be mitigated by selecting appropriate runtime parameters for the Java Virtual Machine (JVM). This entails to exploring parameter space in an exhaustive manner that is not practically feasible for rapid application development. This paper aims at performance enhancement of XML parsing through selection of optimal set of JVM runtime parameters. The proposed approach works independent of parser design. It reduces JVM parameter space through machine learning-based models which are trained using profile data. The impact of parameters is determined using linear regression and artificial neural network-based models. The subsequent computation of a location-based weight vector along with a threshold value for filtration of parameters generates a set of optimal parameters for performance enhancement. The XML parsing code using the optimal parameters achieves average speedups of 13.18% and 21.42% over the standard code on Intel Xeon and Intel Core i7-based systems, respectively.

Transactional Sapphire

Constructing a high-performance garbage collector is hard. Constructing a fully concurrent ‘on-the-fly’ compacting collector is much more so. We describe our experience of implementing the Sapphire algorithm as the first on-the-fly, parallel, replication copying, garbage collector for the Jikes RVM Java virtual machine (JVM). In part, we explain our innovations such as copying with hardware and software transactions, on-the-fly management of Java’s reference types, and simple, yet correct, lock-free management of volatile fields in a replicating collector. We fully evaluate, for the first time, and using realistic benchmarks, Sapphire’s performance and suitability as a low latency collector. An important contribution of this work is a detailed description of our experience of building an on-the-fly copying collector for a complete JVM with some assurance that it is correct. A key aspect of this is model checking of critical components of this complicated and highly concurrent system.

MapOptics: a light-weight, cross-platform visualization tool for optical mapping alignment.

Bionano optical mapping is a technology that can assist in the final stages of genome assembly by lengthening and ordering scaffolds in a draft assembly by aligning the assembly to a genomic map. However, currently, tools for visualization are limited to use on a Windows operating system or are developed initially for visualizing large-scale structural variation. MapOptics is a lightweight cross-platform tool that enables the user to visualize and interact with the alignment of Bionano optical mapping data and can be used for in depth exploration of hybrid scaffolding alignments. It provides a fast, simple alternative to the large optical mapping analysis programs currently available for this area of research. MapOptics is implemented in Java 1.8 and released under an MIT licence. MapOptics can be downloaded from https://github.com/FadyMohareb/mapoptics and run on any standard desktop computer equipped with a Java Virtual Machine (JVM). Supplementary data are available at Bioinformatics online.

FlinkCL: An OpenCL-Based In-Memory Computing Architecture on Heterogeneous CPU-GPU Clusters for Big Data

Research on in-memory big data management and processing has been prompted by the increase in main memory capacity and the explosion in big data. By offering an efficient in-memory distributed execution model, existing in-memory cluster computing platforms such as Flink and Spark have been proven to be outstanding for processing big data. This paper proposes FlinkCL, an in-memory computing architecture on heterogeneous CPU-GPU clusters based on OpenCL that enables Flink to utilize GPU's massive parallel processing ability. Our proposed architecture utilizes four techniques: a heterogeneous distributed abstract model (HDST), a Just-In-Time (JIT) compiling schema, a hierarchical partial reduction (HPR) and a heterogeneous task management strategy. Using FlinkCL, programmers only need to write Java code with simple interfaces. The Java code can be compiled to OpenCL kernels and executed on CPUs and GPUs automatically. In the HDST, a novel memory mapping scheme is proposed to avoid serialization or deserialization between Java Virtual Machine (JVM) objects and OpenCL structs. We have comprehensively evaluated FlinkCL with a set of representative workloads to show its effectiveness. Our results show that FlinkCL improve the performance by up to $11 \times$ for some computationally heavy algorithms and maintains minor performance improvements for a I/O bound algorithm.

Optimizing type-specific instrumentation on the JVM with reflective supertype information

Reflective supertype information (RSI) is useful for many instrumentation-based type-specific analyses on the Java Virtual Machine (JVM). On the one hand, while such information can be obtained when performing the instrumentation within the same JVM process executing the instrumented program, in-process instrumentation severely limits the bytecode coverage of the analysis. On the other hand, performing the instrumentation in a separate process can achieve full bytecode coverage, but complete RSI is generally not available, often requiring the insertion of expensive runtime type checks in the instrumented program. In this article, we present a novel technique to accurately reify complete RSI in a separate instrumentation process. This is challenging, because the observed application may make use of custom classloaders and the loaded classes in one application execution are generally only known upon termination of the application. We implement our technique in an extension of the dynamic analysis framework DiSL. The resulting framework guarantees full bytecode coverage, while providing RSI. Evaluation results on a task profiler demonstrate that our technique can achieve speedups up to a factor of 6.24× wrt. resorting to runtime type checks in the instrumentation code for an analysis with full bytecode coverage.

A multi‐granularity locking scheme for java packedobjects based on a concurrent multiway tree

SummaryIn this paper, we develop a multi‐granularity locking scheme for Java PackedObjects, an experimental enhancement introduced in IBM's J9 Java Virtual Machine. The packed object model organizes data in a multi‐tier manner in which object data can be nested in the container object instead of being pointed to by an object reference, as in the traditional Java object model. This new object data model creates new challenges for multi‐tier data synchronization, requiring concurrent locks on the multi‐tier data of different granularities for maintaining consistency. This is different from the traditional Java synchronization model. In this paper we make use of a concurrent multiway tree to represent the containing and ordering relationship between PackedObjects at different tiers and develop an efficient multi‐granularity locking scheme allowing multiple threads to concurrently manipulate the concurrent multiway tree for synchronization operations. In the evaluation, we compare our new tree‐based multitierSync with the previous multitierSync approaches based on linked‐lists (optimized‐list‐based and lazy‐list‐based). The experimental results show that the tree‐based MultitierPackedSync outperforms the list‐based approaches considerably in different workloads, and the higher the workload, the better the performance gains achieved by the tree‐based MultitierPackedSync.

The design of JVM and native libraries in ScalaLab for efficient scientific computation

ScalaLab is a MATLAB-like environment for the Java Virtual Machine (JVM). ScalaLab is based on the Scala programming language. It utilizes an extensive set of Java and Scala scientific libraries and also has access to many native C/C[Formula: see text] scientific libraries by using mainly the Java Native Interface (JNI).The performance of the JVM platform is continuously improved at a fast pace. Today JVM can effectively support demanding high-performance computing and scales well on multicore platforms. However, sometimes optimized native C/[Formula: see text] code can yield even better performance, by exploiting low-level programming issues, such as optimization of caches and architecture-dependent instruction sets.The present work reports some of the experiences that we gained with experiments with both Just in Time (JIT) JVM code and native code. We compare some aspects of Scala and C[Formula: see text] that concern the requirements of scientific computing and highlight some strong features of the Scala language that facilitate the implementation of scientific scripting. This paper describes how ScalaLab tries to combine the best features of the JVM with those of the C/C[Formula: see text] technology, in order to implement an effective scientific computing environment.

RIM4J: An Architecture for Language-Supported Runtime Measurement against Malicious Bytecode in Cloud Computing

While cloud customers can benefit from migrating applications to the cloud, they are concerned about the security of the hosted applications. This is complicated by the customers not knowing whether their cloud applications are working as expected. Although memory-safety Java Virtual Machine (JVM) can alleviate their anxiety due to the control flow integrity, their applications are prone to a violation of bytecode integrity. The analysis of some Java exploits indicates that the violation results primarily from the given excess sandbox permission, loading flaws in Java class libraries and third-party middlewares and the abuse of sun.misc.UnsafeAPI. To such an end, we design an architecture, called RIM4J, to enforce a runtime integrity measurement of Java bytecode within a cloud system, with the ability to attest this to a cloud customer in an unforgeable manner. Our RIM4J architecture is portable, such that it can be quickly deployed and adopted for real-world purposes, without requiring modifications to the underlying systems and access to application source code. Moreover, our RIM4J architecture is the first to measure dynamically-generated bytecode. We apply our runtime measurement architecture to a messaging server application where we show how RIM4J can detect undesirable behaviors, such as uploading arbitrary files and remote code execution. This paper also reports the experimental evaluation of a RIM4J prototype using both a macro- and a micro-benchmark; the experimental results indicate that RIM4J is a practical solution for real-world applications.

Dual-pivot and beyond: The potential of multiway partitioning in quicksort

Abstract Since 2011 the Java runtime library uses a Quicksort variant with two pivot elements. For reasons that remained unclear for years it is faster than the previous Quicksort implementation by more than 10 %; this is not only surprising because the previous code was highly-tuned and is used in many programming libraries, but also since earlier theoretical investigations suggested that using several pivots in Quicksort is not helpful. In my dissertation I proved by a comprehensive mathematical analysis of all sensible Quicksort partitioning variants that (a) indeed there is hardly any advantage to be gained from multiway partitioning in terms of the number of comparisons (and more generally in terms of CPU costs), but (b) multiway partitioning does significantly reduce the amount of data to be moved between CPU and main memory. Moreover, this more efficient use of the memory hierarchy is not achieved by any of the other well-known optimizations of Quicksort, but only through the use of several pivots.

Dynamic vertical memory scalability for OpenJDK cloud applications

The cloud is an increasingly popular platform to deploy applications as it lets cloud users to provide resources to their applications as needed. Furthermore, cloud providers are now starting to offer a "pay-as-you-use" model in which users are only charged for the resources that are really used instead of paying for a statically sized instance. This new model allows cloud users to save money, and cloud providers to better utilize their hardware. However, applications running on top of runtime environments such as the Java Virtual Machine (JVM) cannot benefit from this new model because they cannot dynamically adapt the amount of used resources at runtime. In particular, if an application needs more memory than what was initially predicted at launch time, the JVM will not allow the application to grow its memory beyond the maximum value defined at launch time. In addition, the JVM will hold memory that is no longer being used by the application. This lack of dynamic vertical scalability completely prevents the benefits of the "pay-as-you-use" model, and forces users to over-provision resources, and to lose money on unused resources. We propose a new JVM heap sizing strategy that allows the JVM to dynamically scale its memory utilization according to the application's needs. First, we provide a configurable limit on how much the application can grow its memory. This limit is dynamic and can be changed at runtime, as opposed to the current static limit that can only be set at launch time. Second, we adapt current Garbage Collection policies that control how much the heap can grow and shrink to better fit what is currently being used by the application. The proposed solution is implemented in the OpenJDK 9 HotSpot JVM, the new release of OpenJDK. Changes were also introduced inside the Parallel Scavenge collector and the Garbage First collector (the new by-default collector in HotSpot). Evaluation experiments using real workloads and data show that, with negligible throughput and memory overhead, dynamic vertical memory scalability can be achieved. This allows users to save significant amounts of money by not paying for unused resources, and cloud providers to better utilize their physical machines.

Evaluating the Java Native Interface (JNI)

This article aims to explore JNI features and to discover fundamental operations of the Java programming language, such as arrays, objects, classes, threads and exception handling, and to illustrate these by using various algorithms and code samples. The authors also investigate the JNI Invocation API that allows native applications to interact with the Java Virtual Machine (JVM). They focus on attaching native threads to a running JVM and on leveraging existing native code using one-to-one mapping and shared stubs.

Espresso

Fast, byte-addressable non-volatile memory (NVM) embraces both near-DRAM latency and disk-like persistence, which has generated considerable interests to revolutionize system software stack and programming models. However, it is less understood how NVM can be combined with managed runtime like Java virtual machine (JVM) to ease persistence management. This paper proposes Espresso, a holistic extension to Java and its runtime, to enable Java programmers to exploit NVM for persistence management with high performance. Espresso first provides a general persistent heap design called Persistent Java Heap (PJH) to manage persistent data as normal Java objects. The heap is then strengthened with a recoverable mechanism to provide crash consistency for heap metadata. Espresso further provides a new abstraction called Persistent Java Object (PJO) to provide an easy-to-use but safe persistence programming model for programmers to persist application data. Evaluation confirms that Espresso significantly outperforms state-of-art NVM support for Java (i.e., JPA and PCJ) while being compatible to data structures in existing Java programs.

Eliminating object reference checks by escape analysis on real-time Java virtual machine

The real-time specification for Java (RTSJ) makes Java to be a real-time programming language. However, the RTSJ’s memory management system is more complicated than J2SE’s. The assignment rules of RTSJ, which prevent the creation of dangling references, must be checked by real-time Java virtual machines (JVMs) at run-time. These frequent run-time object reference checks introduce significant time overheads and unpredictable execution time, which has great impact on real-time systems. This paper presents an equivalence class based, context sensitive and flow insensitive escape analysis algorithm that effectively eliminates unnecessary run-time reference checkpoints of RTSJ programs. The optimization framework has been implemented in an open-source real-time JVM namely jRate and evaluated by CD$$_x$$, a relative authority real-time Java benchmark suite. The results show that this optimization algorithm eliminates more than 90% static reference checkpoints, removes about 50% run-time reference checkpoints on average , and improves the run-time performance of average 3.13%, max 8.93%.

Guía para la configuración de la Interfaz Nativa de Java

En la actualidad existen diversos lenguajes de programación y es necesario contar con una herramienta que permita la integración entre ellos. Java provee de un framework llamado: interfaz nativa de Java, JNI por sus siglas en inglés; para permitir que se puedan escribir programas en otros lenguajes diferentes a Java y mantener la portabilidad entre todas las plataformas. Permite que el código que se ejecuta en la máquina virtual de Java pueda interactuar con aplicaciones y bibliotecas escritas en otros lenguajes, como C, C++ y ensamblador. Esta guía se basa en un ejemplo básico de una aplicación “Hello Word”, la cual servirá para de manera sencilla, mostrar los pasos a seguir para la configuración de JNI en los entornos de desarrollo involucrados, que para esta guía se utilizó, para Java: Eclipse y para C++: Visual Studio 2010.

JPR

Java PathFinder (JPF) is a backtrackable Java Virtual Machine (JVM), which is implemented in Java and runs on a standard JVM (e.g., Oracle HotSpot). Thus, a JPF developer can use off-the- shelf Java debuggers (e.g., jdb) when debugging code that makes up JPF. JPF explores all non-deterministic executions of a given target program and monitors for property violations. To facilitate debugging of the target program, JPF can capture and replay the execution trace that leads to a property violation. While the deterministic replay is invaluable, the replay with JPF does not allow the developer to attach an off-the-shelf Java debugger to the target program (e.g., step through the application code, set breakpoints, etc.). We present a technique, dubbed JPR, to improve the debugging experience of the JPF captured traces by migrating the JPF traces to a new format that can be executed using the standard JVM. JPR annotates each JPF trace, during the capture phase, with ex- tra data (e.g., instruction index, instruction count, etc.); the an- notated trace is then used to instrument Java bytecode to enforce the same execution trace on a standard JVM. JPR is compatible with various optimizations, e.g., state matching and partial-order reduction. We evaluated JPR on all multi-threaded Java pro- grams in the official JPF distribution. Our results show that JPR successfully replayed all JPF traces on the standard JVM with reasonable overhead during both recording and replaying.

Високопродуктивний пакетний добуток матриць афінних перетворень за допомогою Android NDK та JNI

This work contains description of high-performance batched affine matrices multiplication approach in scope of An-droid NDK and JNI. Approach is based on combination of separate technics and practices. Android Native Development Kit and Java Native Interfaces were used to develop and run high-performance code via Java Virtual Machine. This approach overhead was minimized with input data batching for computation. Compile-time optimization approaches were used to speed-up C/C++/Java based solutions. As a development, analysis and research result precise performance benchmarks were made. Output solution was compared with analogues, which are using limited sets of applied techniques, features and approaches. Charts were built on top of that results. Appropriate decisions were made. Problems in programming 2018; 1: 36-45

Cognitive Architecture for a Connected World

We propose a unified architecture for next generation cognitive, low cost, mobile internet. The end user platform is able to scale as per the application and network requirements. It takes computing out of the data center and into end user platform. Internet enables open standards, accessible computing and applications programmability on a commodity platform. The architecture is a super-set to present day infrastructure web computing. The Java virtual machine (JVM) derives from the stack architecture. Applications can be developed and deployed on a multitude of host platforms. O(1) -> O(N). Computing and the internet today are more accessible and available to the larger community. Machine learning has made extensive advances with the availability of modern computing. It is used widely in NLP, Computer Vision, Deep learning and AI. A prototype device for mobile could contain N compute and N MB of memory.

Database-Inspired Optimizations for Statistical Analysis

Computing complex statistics on large amounts of data is no longer a corner case, but a daily challenge. However, current tools such as GNU R were not built to efficiently handle large data sets. We propose to vastly improve the execution of R scripts by interpreting them as a declaration of intent rather than an imperative order set in stone. This allows us to apply optimization techniques from the columnar data management research field. We have implemented several of these optimizers in Renjin, an open-source execution environment for R scripts targeted at the Java virtual machine. The demonstration of our approach using a series of micro-benchmarks and experiments on complex survey analysis show orders-of-magnitude improvements in analysis cost.

Why is garbage collection causing my service level objectives to fail?

Cloud computing abstracts resources and provides them as-a-service to its tenant clients. Platform as a service clouds, which are one of the main types of cloud computing, provide large parts of the hardware/software stack to their users. Cloud systems are expected to abide by certain service level objectives and maintain a certain quality of service, which can be impacted by garbage collection (GC). However, cloud benchmarking is mostly focused in the interconnectivity of cloud services and often neglects the inner workings of language runtimes. In this paper, we present and evaluate CloudGC, a benchmark aiming to stress the GC component of a runtime in various and controllable ways. We then deploy our CloudGC on a cloud system to evaluate the SLO satisfaction of the four GC policies of the IBM J9 Java runtime. Our findings indicate that the default policy Gencon generally outperforms the other three policies, including Balanced, the policy which aims in amortising the costs.