1,721,035 research outputs found
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Full text linkWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Dispelling the Myths Behind First-author Citation Counts
Full text linkWe conducted a full-scale evaluative citation analysis study of scholars in the XML research field to explore just how different from each other author rankings resulting from different citation counting methods actually are, and to demonstrate the capability of emerging data and tools on the Web in supporting more realistic citation counting methods. Our results contest some common arguments for the continued
use of first-author citation counts in the evaluation of scholars, such as high correlations between author rankings by first-author citation counts and other citation
counting methods, and high costs of using more realistic citation counting methods that are not well-supported by the ISI databases. It is argued that increasingly available digital full text research papers make it possible for citation analysis studies to go beyond what the ISI databases have directly supported and to employ more
sophisticated methods
Language Support for Programming High-Performance Code
Full text linkNowadays, the computing landscape is becoming increasingly heterogeneous and this trend is currently showing no signs of turning around. In particular, hardware becomes more and more specialized and exhibits different forms of parallelism. For performance-critical codes it is indispensable to address hardware-specific peculiarities. Because of the halting problem, however, it is unrealistic to assume that a program implemented in a general-purpose programming language can be fully automatically compiled to such specialized hardware while still delivering peak performance. One form of parallelism is single instruction, multiple data (SIMD). Part I of this thesis presents Sierra: an extension for C ++ that facilitates portable and effective SIMD programming. Part II discusses AnyDSL. This framework allows to embed a so-called domain-specific language (DSL) into a host language. On the one hand, a DSL offers the application developer a convenient interface; on the other hand, a DSL can perform domain-specific optimizations and effectively map DSL constructs to various architectures. In order to implement a DSL, one usually has to write or modify a compiler. With AnyDSL though, the DSL constructs are directly implemented in the host language while a partial evaluator removes any abstractions that are required in the implementation of the DSL.Die Rechnerlandschaft wird heutzutage immer heterogener und derzeit ist keine Trendwende in Sicht. Insbesondere wird die Hardware immer spezialisierter und weist verschiedene Formen der Parallelität auf. Für performante Programme ist es unabdingbar, hardwarespezifische Eigenheiten zu adressieren. Wegen des Halteproblems ist es allerdings unrealistisch anzunehmen, dass ein Programm, das in einer universell einsetzbaren Programmiersprache implementiert ist, vollautomatisch auf solche spezialisierte Hardware übersetzt werden kann und dabei noch Spitzenleistung erzielt. Eine Form der Parallelität ist „single instruction, multiple data (SIMD)“. Teil I dieser Arbeit stellt Sierra vor: eine Erweiterung für C++, die portable und effektive SIMD-Programmierung unterstützt. Teil II behandelt AnyDSL. Dieses Rahmenwerk ermöglicht es, eine sogenannte domänenspezifische Sprache (DSL) in eine Gastsprache einzubetten. Auf der einen Seite bietet eine DSL dem Anwendungsentwickler eine komfortable Schnittstelle; auf der anderen Seiten kann eine DSL domänenspezifische Optimierungen durchführen und DSL-Konstrukte effektiv auf verschiedene Architekturen abbilden. Um eine DSL zu implementieren, muss man gewöhnlich einen Compiler schreiben oder modifizieren. In AnyDSL werden die DSL-Konstrukte jedoch direkt in der Gastsprache implementiert und ein partieller Auswerter entfernt jegliche Abstraktionen, die in der Implementierung der DSL benötigt werden
Vectorization system for unstructured codes with a Data-parallel Compiler IR
Full text linkWith Dennard Scaling coming to an end, Single Instruction Multiple Data (SIMD) offers itself as a way to improve the compute throughput of CPUs. One fundamental technique in SIMD code generators is the vectorization of data-parallel code regions. This has applications in outer-loop vectorization, whole-function vectorization and vectorization of explicitly data-parallel languages. This thesis makes contributions to the reliable vectorization of data-parallel code regions with unstructured, reducible control flow. Reducibility is the case in practice where all control-flow loops have exactly one entry point. We present P-LLVM, a novel, full-featured, intermediate representation for vectorizers that provides a semantics for the code region at every stage of the vectorization pipeline. Partial control-flow linearization is a novel partial if-conversion scheme, an essential technique to vectorize divergent control flow. Different to prior techniques, partial linearization has linear running time, does not insert additional branches or blocks and gives proved guarantees on the control flow retained. Divergence of control induces value divergence at join points in the control-flow graph (CFG). We present a novel control-divergence analysis for directed acyclic graphs with optimal running time and prove that it is correct and precise under common static assumptions. We extend this technique to obtain a quadratic-time, control-divergence analysis for arbitrary reducible CFGs. For this analysis, we show on a range of realistic examples how earlier approaches are either less precise or incorrect. We present a feature-complete divergence analysis for P-LLVM programs. The analysis is the first to analyze stack-allocated objects in an unstructured control setting. Finally, we generalize single-dimensional vectorization of outer loops to multi-dimensional tensorization of loop nests. SIMD targets benefit from tensorization through more opportunities for re-use of loaded values and more efficient memory access behavior. The techniques were implemented in the Region Vectorizer (RV) for vectorization and TensorRV for loop-nest tensorization. Our evaluation validates that the general-purpose RV vectorization system matches the performance of more specialized approaches. RV performs on par with the ISPC compiler, which only supports its structured domain-specific language, on a range of tree traversal codes with complex control flow. RV is able to outperform the loop vectorizers of state-of-the-art compilers, as we show for the SPEC2017 nab_s benchmark and the XSBench proxy application.Mit dem Ausreizen des Dennard Scalings erreichen die gewohnten Zuwächse in der skalaren Rechenleistung zusehends ihr Ende. Moderne Prozessoren setzen verstärkt auf parallele Berechnung, um den Rechendurchsatz zu erhöhen. Hierbei spielen SIMD Instruktionen (Single Instruction Multiple Data), die eine Operation gleichzeitig auf mehrere Eingaben anwenden, eine zentrale Rolle. Eine fundamentale Technik, um SIMD Programmcode zu erzeugen, ist der Einsatz datenparalleler Vektorisierung. Diese unterliegt populären Verfahren, wie der Vektorisierung äußerer Schleifen, der Vektorisierung gesamter Funktionen bis hin zu explizit datenparallelen Programmiersprachen. Der Beitrag der vorliegenden Arbeit besteht darin, ein zuverlässiges Vektorisierungssystem für datenparallelen Code mit reduziblem Steuerfluss zu entwickeln. Diese Anforderung ist für alle Steuerflussgraphen erfüllt, deren Schleifen nur einen Eingang haben, was in der Praxis der Fall ist. Wir präsentieren P-LLVM, eine ausdrucksstarke Zwischendarstellung für Vektorisierer, welche dem Programm in jedem Stadium der Transformation von datenparallelem Code zu SIMD Code eine definierte Semantik verleiht. Partielle Steuerfluss-Linearisierung ist ein neuer Algorithmus zur If-Conversion, welcher Sprünge erhalten kann. Anders als existierende Verfahren hat Partielle Linearisierung eine lineare Laufzeit und fügt keine neuen Sprünge oder Blöcke ein. Wir zeigen Kriterien, unter denen der Algorithmus Steuerfluss erhält, und beweisen diese. Steuerflussdivergenz induziert Divergenz an Punkten zusammenfließenden Steuerflusses. Wir stellen eine neue Steuerflussdivergenzanalyse für azyklische Graphen mit optimaler Laufzeit vor und beweisen deren Korrektheit und Präzision. Wir verallgemeinern die Technik zu einem Algorithmus mit quadratischer Laufzeit für beliebiege, reduzible Steuerflussgraphen. Eine Studie auf realistischen Beispielgraphen zeigt, dass vergleichbare Techniken entweder weniger präsize sind oder falsche Ergebnisse liefern. Ebenfalls präsentieren wir eine Divergenzanalyse für P-LLVM Programme. Diese Analyse ist die erste Divergenzanalyse, welche Divergenz in stapelallokierten Objekten unter unstrukturiertem Steuerfluss analysiert. Schließlich generalisieren wir die eindimensionale Vektorisierung von äußeren Schleifen zur multidimensionalen Tensorisierung von Schleifennestern. Tensorisierung eröffnet für SIMD Prozessoren mehr Möglichkeiten, bereits geladene Werte wiederzuverwenden und das Speicherzugriffsverhalten des Programms zu optimieren, als dies mit Vektorisierung der Fall ist. Die vorgestellten Techniken wurden in den Region Vectorizer (RV) für Vektorisierung und TensorRV für die Tensorisierung von Schleifennestern implementiert. Wir zeigen auf einer Reihe von steuerflusslastigen Programmen für die Traversierung von Baumdatenstrukturen, dass RV das gleiche Niveau erreicht wie der ISPC Compiler, welcher nur seine strukturierte Eingabesprache verarbeiten kann. RV kann schnellere SIMD-Programme erzeugen als die Schleifenvektorisierer in aktuellen Industriecompilern. Dies demonstrieren wir mit dem nab_s benchmark aus der SPEC2017 Benchmarksuite und der XSBench Proxy-Anwendung
Applicable and sound polyhedral optimization of low-level programs
Full text linkComputers become increasingly complex. Current and future systems feature configurable hardware, multiple cores with different capabilities, as well as accelerators. In addition, the memory subsystem becomes diversified too. The cache hierarchy grows deeper, is augmented with scratchpads, low-latency memory, and high-bandwidth memory. The programmer alone cannot utilize this enormous potential. Compilers have to provide insight into the program behavior, or even arrange computations and data themselves. Either way, they need a more holistic view of the program. Local transformations, which treat the iteration order, computation unit, and data layout as fixed, will not be able to fully utilize a diverse system. The polyhedral model, a high-level program representation and transformation framework, has shown great success tackling various problems in the context of diverse systems. While it is widely acknowledged for its analytical powers and transformation capabilities, it is also widely assumed to be too restrictive and fragile for real-world programs. In this thesis we improve the applicability and profitability of polyhedral-model-based techniques. Our efforts guarantee a sound polyhedral representation and extend the applicability to a wider range of programs. In addition, we introduce new applications to utilize the information available in the polyhedral program representation, including standalone optimizations and techniques to derive high-level properties
koamabayili/VECTRON-author-checklist: VECTRON author checklist
No full textWe have done our best to complete the author checklist relating to the use of animals in the hut study. Note that the objective for the hut study was to evaluate the IRS treatment applications for residual efficacy against Anopheles mosquitoes, including the local An. coluzzii mosquito population. Cows were only used to attract mosquitoes into the huts and no tests were carried out directly on the cows. The author checklist is intended for use with studies where experiments are carried out on animals, which is why we have had such difficulty in completing this for the hut study, as many of the questions do not relate to how the cows were used
- …
