Dynamic algorithm selection for runtime concepts

Abstract

A key benefit of generic programming is its support for producing modules with clean separation. In particular, generic algorithms are written to work with a wide variety of types without requiring modifications to them. The Runtime concept idiom extends this support by allowing unmodified concrete types to behave in a runtime polymorphic manner. In this paper, we describe one implementation of the runtime concept idiom, in the domain of the C++ standard template library (STL). We complement the runtime concept idiom with an algorithm library that considers both type and concept information to maximize performance when selecting algorithm implementations. We present two implementations, one in ISO C++ and one using an experimental language extension. We use our implementations to describe and measure the performance of runtime-polymorphic analogs of several STL algorithms. The tests demonstrate the effects of different compile-time vs. run-time algorithm selection choices.

C++ lambda expressions and closures

Abstract

A style of programming that uses higher-order functions has become common in C++, following the introduction of the Standard Template Library (STL) into the standard library. In addition to their utility as arguments to STL algorithms, function parameters are useful as callbacks on GUI events, defining tasks to be executed in a thread, and so forth. C++’s mechanisms for defining functions or function objects are, however, rather verbose, and they often force the function’s definition to be placed far from its use. As a result, C++ frustrates programmers in taking full advantage of its own standard libraries. The effective use of modern C++ libraries calls for a concise mechanism for defining small one-off functions in the language, a need that can be fulfilled with lambda expressions.
This paper describes a design and implementation of language support for lambda expressions in C++. C++’s compilation model, where activation records are maintained in a stack, and the lack of automatic object lifetime management make safe lambda functions and closures challenging: if a closure outlives its scope of definition, references stored in a closure dangle. Our design is careful to balance between conciseness of syntax and explicit annotations to guarantee safety. The presented design is included in the draft specification of the forthcoming major revision of the ISO C++ standard, dubbed C++0x. In rewriting typical C++ programs to take advantage of lambda functions, we observed clear benefits, such as reduced code size and improved clarity.

The two paradigms of software development research

Abstract

The most profound conflict in software engineering is not between positivist and interpretivist research approaches or Agile and Heavyweight software development methods, but between the Rational and Empirical Design Paradigms. The Rational and Empirical Paradigms are disparate constellations of beliefs about how software is and should be created. The Rational Paradigm remains dominant in software engineering research, standards and curricula despite being contradicted by decades of empirical research. The Rational Paradigm views analysis, design and programming as separate activities despite empirical research showing that they are simultaneous and inextricably interconnected. The Rational Paradigm views developers as executing plans despite empirical research showing that plans are a weak resource for informing situated action. The Rational Paradigm views success in terms of the Project Triangle (scope, time, cost and quality) despite empirical researching showing that the Project Triangle omits critical dimensions of success. The Rational Paradigm assumes that analysts elicit requirements despite empirical research showing that analysts and stakeholders co-construct preferences. The Rational Paradigm views professionals as using software development methods despite empirical research showing that methods are rarely used, very rarely used as intended, and typically weak resources for informing situated action. This article therefore elucidates the Empirical Design Paradigm, an alternative view of software development more consistent with empirical evidence. Embracing the Empirical Paradigm is crucial for retaining scientific legitimacy, solving numerous practical problems and improving software engineering education.

Highlights

•

Software development research is divided into two incommensurable paradigms.

•

The Rational Paradigm emphasizes problem solving, planning and methods.

•

The Empirical Paradigm emphasizes problem framing, improvisation and practices.

•

The Empirical Paradigm is based on data and science; the Rational Paradigm is based on assumptions and opinions.

•

The Rational Paradigm undermines the credibility of the software engineering research community.

Java in the High Performance Computing arena: Research, practice and experience

Abstract

The rising interest in Java for High Performance Computing (HPC) is based on the appealing features of this language for programming multi-core cluster architectures, particularly the built-in networking and multithreading support, and the continuous increase in Java Virtual Machine (JVM) performance. However, its adoption in this area is being delayed by the lack of analysis of the existing programming options in Java for HPC and thorough and up-to-date evaluations of their performance, as well as the unawareness on current research projects in this field, whose solutions are needed in order to boost the embracement of Java in HPC.

This paper analyzes the current state of Java for HPC, both for shared and distributed memory programming, presents related research projects, and finally, evaluates the performance of current Java HPC solutions and research developments on two shared memory environments and two InfiniBand multi-core clusters. The main conclusions are that: (1) the significant interest in Java for HPC has led to the development of numerous projects, although usually quite modest, which may have prevented a higher development of Java in this field; (2) Java can achieve almost similar performance to natively compiled languages, both for sequential and parallel applications, being an alternative for HPC programming; (3) the recent advances in the efficient support of Java communications on shared memory and low-latency networks are bridging the gap between Java and natively compiled applications in HPC. Thus, the good prospects of Java in this area are attracting the attention of both industry and academia, which can take significant advantage of Java adoption in HPC.

Highlights

► Java is an emerging option for High Performance Computing on multi-core clusters. ► Java can achieve almost similar performance to natively compiled languages. ► Current state of Java for HPC, both for shared and distributed memory programming. ► Performance evaluation of Java message-passing and Java threads on multi-core clusters. ► Java is an alternative to HPC programming as it obtains comparable scalability results.