Janice E. Cuny

Portable profiling and tracing for parallel, scientific applications using C++

1. Abstract Performance measurement of parallel, objectoriented (00) programs requires the development of instrumentation and analysis techniques beyond those used for more traditional languages. Performance events must be redefined for the conceptual 00 programming model, and those events must be instrumented and tracked in the context of 00 language abstractions, compilation methods, and runtime execution dynamics. In this paper, we focus on the profiling and tracing of C++ applications that have been written using a rich parallel programming framework for highperformance, scientific computing. We address issues of class-based profiling, instrumentation of templates, runtime function identification, and polymorphic (type-based) profiling. Our solutions are implemented in the TAU portable profiling package which also provides support for profiling groups and userlevel timers. We demonstrate TAU’s C++ profiling capabilities for real parallel applications, built from components of the ACTS toolkit. Future directions include work on runtime performance data access, dynamic instrumentation, and higher-level performance data analysis and visualization that relates object semantics with performance execution behavior.

A Tool Framework for Static and Dynamic Analysis of Object-Oriented Software with Templates

The developers of high-performance scientific applications often work in complex computing environments that place heavy demands on program analysis tools. The developers need tools that interoperate, are portable across machine architectures, and provide source-level feedback. In this paper, we describe a tool framework, the Program Database Toolkit (PDT), that supports the development of program analysis tools meeting these requirements. PDT uses compile-time information to create a complete database of high-level program information that is structured for well-defined and uniform access by tools and applications. PDT’s current applications make heavy use of advanced features of C++, in particular, templates. We describe the toolkit, focussing on its most important contribution -- its handling of templates -- as well as its use in existing applications.

Recruitment and retention of women graduate students in computer science and engineering: results of a workshop organized by the computing research association

T his document is the report of a workshop that convened a group of experts to discuss the recruitment and retention of women in Computer Science and Engineering (CSE) Graduate Programs. (1) Participants included long-time members of the CSE academic and research communities, social scientists engaged in relevant research, and directors of successful retention efforts. (2) The report is a compendium of the experience and expertise of workshop participants, rather than the result of a full-scale, scholarly study into the range of issues. Its goal is to provide departments with practical advice on recruitment and retention in the form of a set of specific recommendations. Women are significantly underrepresented in CSE academic departments [5][40]. As computing technology becomes increasingly pervasive, this underrepresentation translates into a loss of opportunity for individuals, a loss of talent to the workforce, and a loss of creativity in shaping the future of society. While there are many causes of this underrepresentation some rooted in early socialization and primary educational experiences academic departments at the university level nevertheless can have an effect [6][7]. In particular, an improvement at the graduate level in recruitment and retention (and thus in graduation rates) would enable more women to move into visible and influential positions in the CSE community. The increasing presence of these women would provide positive role models and mentors. In order to treat all students fairly, educators must pay attention to gender-based traits. Although in characterizing behaviors one must be careful to acknowledge the existence of individual differences and to avoid stereotyping, there is a large body of information on gender traits. There is strong evidence, for example, that women, even though they perform at the same levels, have less confidence in their abilities and individual accomplishments than men [2][17][36][39][43]. Women are often less aggressive than male students in promoting themselves, attempting new or challenging activities, and pursuing awards or fellowships. There is evidence that females come to computing as only one interest among many, and are thus less single-minded than their male counterparts [27]. Often women report feeling out of place in the male-dominated, hacker culture [3] [22] [28]. In light of such differences, some of our recommendations are gender-specific. Most, however, are not. The adoption of our recommendations would improve the educational environment for all students. The recommendations are given in two sections, the first

Building Domain-Specific Environments for Computational Science: a Case Study in Seismic Tomography

We report on our experiences in building a computational environment for tomographic image analysis for marine seismologists studying the structure and evolution of mid- ocean ridge volcanism. The computational environment is determined by an evolving set of requirements for this problem domain and includes needs for high performance parallel computing, large data analysis, model visualiza tion, and computation interaction and control. Although these needs are not unique in scientific computing, the integration of techniques for seismic tomography with tools for parallel computing and data analysis into a com putational environment was (and continues to be) an interesting, important learning experience for researchers in both disciplines. For the geologists, the use of the environment led to fundamental geologic discoveries on the East Pacific Rise, the improvement of parallel ray-trac ing algorithms, and a better regard for the use of compu tational steering in aiding model convergence. The com puter scientists received valuable feedback on the use of programming, analysis, and visualization tools in the en vironment. In particular, the tools for parallel program data query (DAQV) and visualization programming (Viz) were demonstrated to be highly adaptable to the problem do main. We discuss the requirements and the components of the environment in detail. Both accomplishments and limitations of our work are presented.

Event and state-based debugging in TAU: a prototype

Parallel programs are complex and often require a multilevel debugging strategy that combines both eventand state-based debugging. We report here on preliminary work that combines these approaches within the TAU program analysis environment for pC++. This work extends the use of event-based modeling to object-parallel languages, provides an alternative mechanism for establishing meaningful global breakpoints in object-oriented languages, introduces the TAU program interaction and control infrastructure, and provides an environment for the assessment of mixed eventand state-based strategies.

The potential coupling interface: metadata for model coupling

Model coupling is a nontrivial task that is not adequately supported in existing frameworks. Our long term goal is to support the fast-prototyping of model couplings, enabling scientists to quickly experiment with a variety of linkings without having to make an upfront investment in reprogramming. This paper introduces the centerpiece of our framework, the potential coupling interface (PCI), a visual representation of a model code based on simplified control flow graphs. The PCI serves three roles: it is a new form of metadata describing the coupling potential of a model; it is the vehicle for the specification of couplings; and it is the basis for automatic code generation. It is easy to specify and once specified, it is available for all future coupling activities. The PCI allows scientists to focus on the important domain and model issues of coupling without having to revisit legacy code for each new effort.

Why Web GIS May Not Be Enough: A Case Study with the Virtual Research Vessel

During several decades of investigation, the East Pacific Rise seafloor-spreading center at 9°-10°N has been explored by marine geologists, geophysicists, chemists, and biologists, and has emerged as one of the best studied sections of the global midocean ridge. It is an example of a region for which there is now a great wealth of observational data, results, and data-driven theoretical studies. However, these have yet to be fully utilized, either by research scientists or educators. While the situation is improving, a large amount of data, results, and related theoretical models still exist either in an inert, noninteractive form (e.g., journal publications) or as unlinked and currently incompatible computer data or algorithms. Presented here is the prototype of a computational environment and toolset, called the Virtual Research Vessel, to improve the situation by providing marine scientists and educators with simultaneous access to data, maps, and numerical models. While infrastructure is desired and needed for ready access to data and the resulting maps via web GIS in order to link disparate data sets (data to data), it is argued that data must also be linked to models for better exploration of new relations between observables, refinement of numerical simulations, and the quantitative evaluation of scientific hypotheses. For widespread data access, web GIS is therefore only a preliminary step rather than a final solution, and the ongoing implementation of the Virtual Research Vessel (scheduled for final completion in 2004-2005) is a case study for the midocean ridge community to test the effectiveness of moving beyond the data-to-data mode towards data-to-models and data-to-interpretation.

A domain-specific language for model coupling

There is an increasing need for the comprehensive simulation of complex, dynamic, physical systems. Often such simulations are built by coupling existing, component models so that their concurrent simulations affect each other. The process of model coupling is, however, a nontrivial task that is not adequately supported by existing frameworks. To provide better support, we have developed an approach to model coupling that uses high level model interfaces called potential coupling interfaces. In this work, we present a visual, domain-specific language for model coupling, called the coupling description language, based on these interfaces. We show that it supports the resolution of model incompatibilities and allows for the fast-prototyping of coupled models

An IL converter and program database for analysis tools

1. ABSTRACT Developers of static and dynamic analysis tools for C++ programs need access to information on functions, classes, templates, and macros in parsed C++ code. Existing tools, such as the EDG display tool, provide that access, but in an unsuitable format. We built a converter that prunes and reorganizes the information into the appropriate format. The converter provides the information needed for our TAU (Tuning and Analysis Utilities) tools and, in more general terms, provides C++ developers considerable opportunities for automating software development.

A Prototype Notebook-Based Environment for Computational Tools Computational Tools

The Virtual Notebook Environment (ViNE) is a platform-independent, web-based interface designed to support a range of scientific activities across distributed, heterogeneous computing platforms. ViNE provides scientists with a web-based version of the common paper-based lab notebook, but in addition, it provides support for collaboration and management of computational experiments. Collaboration is supported with the web-based approach, which makes notebook material generally accessible and with a hierarchy of security mechanisms that screen that access. ViNE provides uniform, system-transparent access to data, tools, and programs throughout the scientists computing infrastructure. Computational experiments can be launched from ViNE using a visual specification language. The scientist is freed from concerns about inter-tool connectivity, data distribution, or data management details. ViNE also provides support for dynamically linking analysis results back into the notebook content. In this paper we present the ViNE system architecture and a case study of its use in neuropsychology research at the University of Oregon. Our case study with the Brain Electrophysiology Laboratory (BEL) addresses their need for data security and management, collaborative support, and distributed analysis processes. The current version of ViNE is a prototype system being tested with this and other scientific applications.