Search

This work is aimed at creating an automatically parallelizing compiler for a distributed memory computing system. The conditions for the correctness of parallelization of the considered sequential programs on a computing system with distributed memory are described. The article considers the problem of placing data in distributed memory with minimization of interprocessor transfers for a program represented by a sequence of parallelizable loops. To solve this problem, an auxiliary bipartite "statements-variables" graph (SVG) is constructed and analyzed from the text of the program. The problem of grouping transfers of small data volumes into a smaller number of transfers of large data volumes is considered. The solution of this problem also leads to minimization of the time for data transfer. New chips with distributed local memory of thousands of cores have higher performance than the previous ones. Therefore, the development of compilers for the development of such microcircuits is necessary and relevant. [ABSTRACT FROM AUTHOR]

When talking about digital transformation, data sovereignty considerations and data transfers cannot be excluded from the discussion, given the considerable likelihood that digital technologies deployed along the process collect, process and transfer (personal) data in multiple jurisdictions. An increasing number of nations, especially those within the BRICS grouping (Brazil, Russia, India, China, and South Africa) are developing their data governance and digital transformation approaches based on data sovereignty considerations, deeming specific types of data as key strategic and economic resources, which deserve particular protection and that must be leveraged for national development. From this perspective, this paper will try to shed light on how data sovereignty and data transfers interplay in the context of digital transformations. Particularly, we will consider the various dimensions that compose the concept of data sovereignty and will utilise a range of examples from the BRICS grouping to back some of the key considerations developed with empirical evidence. We define data sovereignty as the capacity to understand how and why (personal) data are processed and by whom, develop data processing capabilities, and effectively regulate data processing, thus retaining self-determination and control. We have chosen the BRICS grouping for three reasons. First, research on the grouping's data policies and digital transformation is still minimal despite their leading role. Second, BRICS account for over 40 % of the global population, or 3.2 billion people (which can be seen as 3.2 billion "data subjects" or data producers, depending on perspective, thus making them key players in data governance and digital transformation. Third, the BRICS members have realised that digital transformation is essential for the future of their economies and societies and have shaped specific data governance visions which must be considered by other countries, especially from the global majority, to understand why data governance is instrumental to foster thriving digital environments. [ABSTRACT FROM AUTHOR]

It is proposed an algorithm finding the optimal array placement in the distributed memory of a parallel computing system. This work is a step towards the development of a new generation of parallelizing compilers for computing systems with a distributed memory. Such compilers may be useful for manycore systems on chip with addressable local memory (not cache) for each core. This is especially important for systems on chip with a many processor cores, where data exchange with RAM is a performance bottleneck. The construction of special auxiliary program graph for estimation of minimum number of data transfers is obtained. The arrays placement with the optimal number of data transfers is constructing in this paper. [ABSTRACT FROM AUTHOR]

Abstract: With the rapid development of IT data driven experiments and simulations have become very advanced and complicated. The amount and complexity of scientific data increase exponentially. Nowadays the challenge is to efficiently support scientists who generate more data than they can possibly look at and understand. This requires not only high-performance visualization and feature extraction techniques but also efficient and intuitive management of available data and underlying computational resources. Consequently, Service Oriented Architectures and workflow management systems are becoming popular solutions for the deployment of e-Science Infrastructures aimed to assist in exploration and analysis of large scientific data. In this paper, we compare two transport models of workflow execution for data-intensive medical visualizations that rely on web services. The image-based analysis of vascular disorders served as a case study for this project. We applied a service oriented approach to construct distributed visualization pipelines, which allow visualization experts to develop visualizations to view and interact with large medical data sets. Moreover, end-users (i.e., medical specialists) can explore these visualizations irrespective of their geographical location and available computing resources. The paper reports on the current implementation status and presents our main findings. [ABSTRACT FROM AUTHOR]