Search results

The growing gap between processor and memory speeds has lead to complex memory hierarchies as processors evolve to mitigate such divergence by exploiting the locality of reference. In this direction, the BSC performance analysis tools have been recently extended to provide insight into the application memory accesses by depicting their temporal and spatial characteristics, correlating with the source-code and the achieved performance simultaneously. These extensions rely on the Precise Event-Based Sampling (PEBS) mechanism available in recent Intel processors to capture information regarding the application memory accesses. The sampled information is later combined with the Folding technique to represent a detailed temporal evolution of the memory accesses and in conjunction with the achieved performance and the source-code counterpart. The reports generated by the latter tool help not only application developers but also processor architects to understand better how the application behaves and how the system performs. In this paper, we describe a tighter integration of the sampling mechanism into the monitoring package. We also demonstrate the value of the complete workflow by exploring already optimized state–of–the–art benchmarks, providing detailed insight of their memory access behavior. We have taken advantage of this insight to apply small modifications that improve the applications' performance. ; This work has been performed in the Intel-BSC Exascale Lab. We would like to thank Forschungszentrum Jülich for the compute time on the Jureca system. This project has received funding from the European Union's Horizon 2020 research and innovation program under Marie Sklodowska-Curie grant agreement no. 749516. ; Peer Reviewed ; Postprint (author's final draft)

Operating systems have historically had to manage only a single type of memory device. The imminent availability of heterogeneous memory devices based on emerging memory technologies confronts the classic single memory model and opens a new spectrum of possibilities for memory management. Transparent data movement between different memory devices based on access patterns of applications is a desired feature to make optimal use of such devices and to hide the complexity of memory management to the end user. However, capturing memory access patterns of an application at runtime comes at a cost, which is particularly challenging for large-scale parallel applications that may be sensitive to system noise. In this work, we focus on the access pattern profiling phase prior to the actual memory relocation. We study the feasibility of using Intel's Processor Event-Based Sampling (PEBS) feature to record memory accesses by sampling at runtime and study the overhead at scale. We have implemented a custom PEBS driver in the IHK/-McKernel lightweight multi-kernel operating system, one of whose advantages is minimal system interference due to the lightweight kernel's simple design compared to other OS kernels such as Linux. We present the PEBS overhead of a set of scientific applications and show the access patterns identified in noise sensitive HPC applications. Our results show that clear access patterns can be captured with a 10% overhead in the worst-case and 1% in the best case when running on up to 128k CPU cores (2,048 Intel Xeon Phi Knights Landing nodes). We conclude that online memory access profiling using PEBS at large-scale is promising for memory management in heterogeneous memory environments. ; This work has been partially funded by MEXT's program for the Development and Improvement of Next Generation Ultra High- Speed Computer Systems under its subsidies for operating the Specific Advanced Large Research Facilities in Japan. This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska- Curie grant agreement No 708566 (DURO). ; Peer Reviewed ; Postprint (author's final draft)

Operating systems have historically had to manage only a single type of memory device. The imminent availability of heterogeneous memory devices based on emerging memory technologies confronts the classic single memory model and opens a new spectrum of possibilities for memory management. Transparent data movement between different memory devices based on access patterns of applications is a desired feature to make optimal use of such devices and to hide the complexity of memory management to the end user. However, capturing memory access patterns of an application at runtime comes at a cost, which is particularly challenging for large-scale parallel applications that may be sensitive to system noise. In this work, we focus on the access pattern profiling phase prior to the actual memory relocation. We study the feasibility of using Intel's Processor Event-Based Sampling (PEBS) feature to record memory accesses by sampling at runtime and study the overhead at scale. We have implemented a custom PEBS driver in the IHK/-McKernel lightweight multi-kernel operating system, one of whose advantages is minimal system interference due to the lightweight kernel's simple design compared to other OS kernels such as Linux. We present the PEBS overhead of a set of scientific applications and show the access patterns identified in noise sensitive HPC applications. Our results show that clear access patterns can be captured with a 10% overhead in the worst-case and 1% in the best case when running on up to 128k CPU cores (2,048 Intel Xeon Phi Knights Landing nodes). We conclude that online memory access profiling using PEBS at large-scale is promising for memory management in heterogeneous memory environments. ; This work has been partially funded by MEXT's program for the Development and Improvement of Next Generation Ultra High- Speed Computer Systems under its subsidies for operating the Specific Advanced Large Research Facilities in Japan. This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska- Curie grant agreement No 708566 (DURO). ; Peer Reviewed ; Postprint (author's final draft)

RESUMO O artigo aborda a importância de preservar e de manter publicamente disponíveis os acervos da comunicação governamental nos sites oficiais, a partir do entendimento de que esses são espaços privilegiados para difusão e pesquisa sobre informações de interesse público nas democracias. Como forma de evidenciar o tema, apresenta levantamento da comunicação governamental disponível no site oficial da Presidência da República do Brasil produzida durante o governo do ex-presidente Michel Temer (2016-2018) e da ex-presidenta Dilma Rousseff (2010-2016), discutindo os impactos da preservação e do apagamento da comunicação governamental para a memória política do país. A abordagem vincula-se ao campo da democracia digital e responde a perspectivas teóricas das humanidades digitais, especificamente sobre registro, apagamento e memória na internet.Palavras-chave: Comunicação Pública; Comunicação Governamental; Democracia Digital; Sites Oficiais; Memória Política.ABSTRACT This paper discusses the importance of preserving and keeping publicly available collections of government communication on official websites, as privileged spaces for dissemination and research on information of public interest in democracies. As a way of highlighting the theme, this article observes the government communication available on the official website of the Brazil presidency, produced during the government of former President Michel Temer (2016-2018) and former President Dilma Rousseff (2010-2016) in order to discuss the impacts of preservation and erasure of government communication on the political memory of the country. The approach is linked to the Digital Democracy and the theoretical perspectives of the Digital Humanities, specially about registration, erasure and political memory on the internet.Keywords: Public Communication; Government Communication; Digital Democracy; Official Websites; Political Memory.

This research tested a transactive theory model of how individuals allocate and retrieve task-related information in work teams. It extended prior research by exploring the role of communal information repositories in the context of human information resources. Structural equation modeling of six integrated hypotheses revealed several significant results. First, usage of information repositories was significantly related to individual access to information. However, the relationship between individual direct information exchange with team members (the human repositories) and individual access to information was significant only among average-level users of organizational information repositories. Second, development of individual expertise directories significantly influenced individual direct information exchange with team members. Third, perceived usage of organizational information repositories by team members significantly influenced actual usage. Finally, technology-specific competence in using intranets significantly influenced the actual usage of intranets as organizational information repositories.

This article reflects about two aspects of information science and, more specifically, of archival science: the access to information and memory. Both aspects are referred to scientific information, that is, that information produced and kept in scientific institutions to support science; and to two specific institutions: the State University of Campinas (UNICAMP) and the São Paulo State University "Júlio de Mesquita Filho" (UNESP). The study of access to scientific information and memory is in the frontier between social and archival studies. The reflection is also one in the frontier, as pretends to establish conceptual relations with other areas of the human sciences as history, sociology and politics. ; Se reflexiona sobre dos aspectos centrales de la ciencia de la información y, más específicamente, de la archivología: el acceso a la información científica y la memoria. Se focaliza la reflexión en la información científica, o sea, aquella producida y guardada en instituciones científicas y para la ciencia, en dos instituciones concretas: la Universidad Estatal de Campinas (UNICAMP) y la Universidad Estatal Paulista "Júlio de Mesquita Filho" (UNESP). El estudio de acceso a la información y de la memoria científica está en lo que podemos denominar la frontera entre los estudios del archivo y de la sociedad. La reflexión también es de frontera, en la medida en que establece relaciones conceptuales con otras áreas de las ciencias humanas como la historia, la sociología y la política.

In some Australian academic circles in the 1980s it was believed that, as the numbers of soldiers of the world wars declined over time, so would attendances at war remembrance ceremonies on Anzac Day and interest in war commemoration in general. Contrary to expectation, however, there has been a steady rise in eagerness for war memory in Australia over the past three decades manifest in media interest and increasing attendance at Anzac Day services. Rather than dying out, 'Anzac' is being reinvented for new generations. Emerging from this phenomenon has been a concomitant rise in war memorial and commemorative landscape building across Australia fuelled by government funding (mostly federal) and our relentless search for a national story. Many more memorial landscapes have been built in Western Australia over the past thirty years than at the end of either of the World Wars, a trend set to peak in 2014 with the Centenary of Anzac. This paper examines the origins and progress of this boom in memorial building in Western Australia and argues that these new memorial settings establish 'circuits of memory' which ultimately re-enchant and reinforce the Anzac renaissance.

Archives, just like museums or libraries, are agents which contribute to the creation of our cultural memory. Inextricably linked to the notion of cultural heritage, they highlight some narratives, while sidelining or excluding others. Therefore it is important to critically reflect on the question "whose heritage" (Stuart Hall) is created in the process of archiving. This article looks at the politics of creating access to audiovisual heritage in European film archives via online video streaming opportunities. Examining audiovisual archives as agents in the construction of transnational memories, my research aims to provide new ways of reflecting on diversity practices in archival selection. As case study, this article examines the archival politics of the national film archives in Sweden, especially the way archivists are curating the site Filmarkivet.se. ; VR-projekt "Den rörliga bildens kulturarv. Mångfald och minne i digitala filmarkiv" (2016-2018)

This paper explores the modern metropolis as an ironically concrete metaphor for the collective memory and the mourning of cinema's passing, as it—the "city"—is digitally constructed in two recent, auteur-directed, special effects-driven blockbuster films, Inception and Hugo. The modern city, and mass media, such as the cinema, as well as modes of mass transport, especially the train, all originate in the 19th century, but come into their own in the early 20th century in their address to a subject as the mobilised citizen-consumer who, as Anne Friedberg makes clear, is also always a viewer. Additionally, as Barbara Mennel has recently shown, the advent in Europe of trains and time zones, in their transformation of modern time and space, paved the way for cinema's comparably cataclysmic impact upon modern subjectivity in its iconic reproduction of movement within illusory 3D space. Both films, thus, in their different ways employ cinematic remediation as a form of cultural memory whose nostalgia for cinema's past is rendered with the latest digital effects, hidden in plain sight in the form of subjective memories (as flashback) and dreams. While a version of this reading has been advanced before (at least for Hugo), this paper goes further by connecting each film's status as remediated dream-memory to its respective dependence upon the city as a post-cinematic three-dimensional framework within which locative and locomotive desires alike determine a subject whose psyche is indistinguishable from the cityscape that surrounds him.

During 20112014, I conducted iterative rounds of field research at state-mandated Historical Memory Institutes that emerged in postsocialist East-Central Europe. The multi-sited fieldwork was focused at the Romanian Institute for the Investigation of Communist Crimes and Memory of the Romanian Exile (IICCMER), at the Czech Institute for the Study of Totalitarian Regimes (??STR), and at the Albanian Institute for the Study of Communist Crimes and Aftereffects (ISKK). The objective of the research was to understand how the emergence of these particular institutions affected processes of addressing the past. This case study problematizes the process of institutional access and immersion in the field as central to the fieldwork experience in allegedly bureaucratic institutions established in politicized settings. The dynamic process of public contestation toward these Institutes, doings and undoings of institutional structures, and hierarchies provides the social context of the intricacy of institutional access. I reflect on the quandaries, dilemmas, strategies, and limitations of field research within various dynamic state institutions. Striving for ideal conditions of complete immersion in the field can be either illusory or counterproductive in a context of competing representations on the past and when the stakes are high. Ethnographic sensibility is an important component of fieldwork. Rather than taking sides, or remaining neutral to various respondents and informants, of opposing positions, the researcher needs to show ethnographic sensibility to each. Central issues that structure this case study focus on institutional access, field immersion, advantages of multi-sited ethnography, strategies of overcoming fieldwork limitations, ambiguity of Institut??s researchers/bureaucrats, and the reflexivity of the ethnographer. Ultimately, generating knowledge and understanding in political science through field research requires relishing and revealing apparent similarities and latent contradictions.

The invasion of multi-core and multi-processor platforms on all aspects of computing makes shared memory parallel programming mainstream. Yet, the fundamental problems of exploiting parallelism efficiently and correctly have not been fully addressed. Moreover, the execution model of these platforms (notably the relaxed memory models they implement) introduces new challenges to static and dynamic program analysis. In this work we address 1) the optimization of pessimistic implementations of critical sections and 2) the dynamic information flow analysis for parallel executions of multi-threaded programs. Critical sections are excerpts of code that must appear as executed atomically. Their pessimistic implementation reposes on synchronization mechanisms, such as mutexes, and consists into obtaining and releasing them at the beginning and end of the critical section respectively. We present a general algorithm for the acquisition/release of synchronization mechanisms and define on top of it several policies aiming to reduce contention by minimizing the possession time of synchronization mechanisms. We demonstrate the correctness of these policies (i.e. they preserve atomicity and guarantee deadlock freedom) and evaluate them experimentally. The second issue tackled is dynamic information flow analysis of parallel executions. Precisely tracking information flow of a parallel execution is infeasible due to non-deterministic accesses to shared memory. Most existing solutions that address this problem enforce a serial execution of the target application. This allows to obtain an explicit serialization of memory accesses but incurs both an execution-time overhead and eliminates the effects of relaxed memory models. In contrast, the technique we propose allows to predict the plausible serializations of a parallel execution with respect to the memory model. We applied this approach in the context of taint analysis , a dynamic information flow analysis widely used in vulnerability detection. To improve precision of taint analysis we further take into account the semantics of synchronization mechanisms such as mutexes, which restricts the predicted serializations accordingly. The solutions proposed have been implemented in proof of concept tools which allowed their evaluation on some hand-crafted examples. ; L'utilisation massive des plateformes multi-cœurs et multi-processeurs a pour effet de favoriser la programmation parallèle à mémoire partagée. Néanmoins, exploiter efficacement et de manière correcte le parallélisme sur ces plateformes reste un problème de recherche ouvert. De plus, leur modèle d'exécution sous-jacent, et notamment les modèles de mémoire "relâchés", posent de nouveaux défis pour les outils d'analyse statiques et dynamiques. Dans cette thèse nous abordons deux aspects importants dans le cadre de la programmation sur plateformes multi-cœurs et multi-processeurs: l'optimisation de sections critiques implémentées selon l'approche pessimiste, et l'analyse dynamique de flots d'informations. Les sections critiques définissent un ensemble d'accès mémoire qui doivent être exécutées de façon atomique. Leur implémentation pessimiste repose sur l'acquisition et le relâchement de mécanismes de synchronisation, tels que les verrous, en début et en fin de sections critiques. Nous présentons un algorithme générique pour l'acquisition/relâchement des mécanismes de synchronisation, et nous définissons sur cet algorithme un ensemble de politiques particulier ayant pour objectif d'augmenter le parallélisme en réduisant le temps de possession des verrous par les différentes threads. Nous montrons alors la correction de ces politiques (respect de l'atomicité et absence de blocages), et nous validons expérimentalement leur intérêt. Le deuxième point abordé est l'analyse dynamique de flot d'information pour des exécutions parallèles. Dans ce type d'analyse, l'enjeu est de définir précisément l'ordre dans lequel les accès à des mémoires partagées peuvent avoir lieu à l'exécution. La plupart des travaux existant sur ce thème se basent sur une exécution sérialisée du programme cible. Ceci permet d'obtenir une sérialisation explicite des accès mémoire mais entraîne un surcoût en temps d'exécution et ignore l'effet des modèles mémoire relâchées. A contrario, la technique que nous proposons permet de prédire l'ensemble des sérialisations possibles vis-a-vis de ce modèle mémoire à partir d'une seule exécution parallèle ("runtime prediction"). Nous avons développé cette approche dans le cadre de l'analyse de teinte, qui est largement utilisée en détection de vulnérabilités. Pour améliorer la précision de cette analyse nous prenons également en compte la sémantique des primitives de synchronisation qui réduisent le nombre de sérialisations valides. Les travaux proposé ont été implémentés dans des outils prototype qui ont permit leur évaluation sur des exemples représentatifs.