13,033 research outputs found
MABS validation through repeated execution and data mining analysis
No full textAgent Based Modelling is the most interesting and advanced approach for simulating a complex system: in a social context, the single parts and the whole are often very hard to describe in detail. Besides, there are agent based formalisms which allow to study the emergency of social behaviour with the creation and study of models, known as artificial societies. Thanks to the ever increasing computational power, it's been possible to use such models to create software, based on intelligent agents, which aggregate behaviour is complex and difficult to predict, and can be used in open and distributed systems. Data mining is born in the last decades in order to help users in finding useful knowledge from the otherwise overwhelming amount of data available nowadays from the web and the data collected every day by companies. Data Mining techniques can therefore be the keystone to reveal non-trivial knowledge expressed by the initial assumption used to build the micro-level of the model and the structure of the society of agents that emerged from the simulation
Accelerating Large-Scale Graph Processing with FPGAs: Lesson Learned and Future Directions
No full textProcessing graphs on a large scale presents a range of difficulties, including irregular memory access patterns, device memory limitations, and the need for effective partitioning in distributed systems, all of which can lead to performance problems on traditional architectures such as CPUs and GPUs. To address these challenges, recent research emphasizes the use of Field-Programmable Gate Arrays (FPGAs) within distributed frameworks, harnessing the power of FPGAs in a distributed environment for accelerated graph processing. This paper examines the effectiveness of a multi-FPGA distributed architecture in combination with a partitioning system to improve data locality and reduce inter-partition communication. Utilizing Hadoop at a higher level, the framework maps the graph to the hardware, efficiently distributing pre-processed data to FPGAs. The FPGA processing engine, integrated into a cluster framework, optimizes data transfers, using offline partitioning for large-scale graph distribution. A first evaluation of the framework is based on the popular PageRank algorithm, which assigns a value to each node in a graph based on its importance. In the realm of large-scale graphs, the single FPGA solution outperformed the GPU solution that were restricted by memory capacity and surpassing CPU speedup by 26x compared to 12x. Moreover, when a single FPGA device was limited due to the size of the graph, our performance model showed that a distributed system with multiple FPGAs could increase performance by around 12x. This highlights the effectiveness of our solution for handling large datasets that surpass on-chip memory restrictions.Quantum Circuit Architectures and Technolog
Distributed large-scale graph processing on FPGAs
No full textAbstract Processing large-scale graphs is challenging due to the nature of the computation that causes irregular memory access patterns. Managing such irregular accesses may cause significant performance degradation on both CPUs and GPUs. Thus, recent research trends propose graph processing acceleration with Field-Programmable Gate Arrays (FPGA). FPGAs are programmable hardware devices that can be fully customised to perform specific tasks in a highly parallel and efficient manner. However, FPGAs have a limited amount of on-chip memory that cannot fit the entire graph. Due to the limited device memory size, data needs to be repeatedly transferred to and from the FPGA on-chip memory, which makes data transfer time dominate over the computation time. A possible way to overcome the FPGA accelerators’ resource limitation is to engage a multi-FPGA distributed architecture and use an efficient partitioning scheme. Such a scheme aims to increase data locality and minimise communication between different partitions. This work proposes an FPGA processing engine that overlaps, hides and customises all data transfers so that the FPGA accelerator is fully utilised. This engine is integrated into a framework for using FPGA clusters and is able to use an offline partitioning method to facilitate the distribution of large-scale graphs. The proposed framework uses Hadoop at a higher level to map a graph to the underlying hardware platform. The higher layer of computation is responsible for gathering the blocks of data that have been pre-processed and stored on the host’s file system and distribute to a lower layer of computation made of FPGAs. We show how graph partitioning combined with an FPGA architecture will lead to high performance, even when the graph has Millions of vertices and Billions of edges. In the case of the PageRank algorithm, widely used for ranking the importance of nodes in a graph, compared to state-of-the-art CPU and GPU solutions, our implementation is the fastest, achieving a speedup of 13 compared to 8 and 3 respectively. Moreover, in the case of the large-scale graphs, the GPU solution fails due to memory limitations while the CPU solution achieves a speedup of 12 compared to the 26x achieved by our FPGA solution. Other state-of-the-art FPGA solutions are 28 times slower than our proposed solution. When the size of a graph limits the performance of a single FPGA device, our performance model shows that using multi-FPGAs in a distributed system can further improve the performance by about 12x. This highlights our implementation efficiency for large datasets not fitting in the on-chip memory of a hardware device
Una crítica a las políticas capitalistas que obstaculizan el libre flujo de documentación y comunicación científica pública general y deportiva
No full textIntroduction: This paper (A critique to the capitalist policies that hamper the free flow of public scientific documentation and communication in general and sports) is an introductory and critical literature review of some of the major capitalistic policies that hamper the free flow of public scientific documentation and communication in general and in the sciences of sports in particular. Aim: The aim of this paper is to make the readers aware to what extent such policies hamper them to access publications and communicate freely, free of charge, democratically, communally, and at the same time to show them free of charge Open Access alternatives. Method: Only hermeneutics and documental methods were employed. Results: These were the major capitalistic policies found in general and in sports in particular: 1) the intervention of capitalistic corporations and bourgeois States within the public universities, and in the public affairs is transforming education, science, and information-communication into an appendix without criticism in favor of capitalistic corporations (Dilevko, 2009; Moles Plaza, 2006; Muela Meza, 2005; Bakan, 2004; Marquand, 2004; Verzola, 2004; González Barbone, 2002; Fox 1983); 2) the buying-selling of information-commmunication financed with public money affects the scant budgets of public universities, and the salaries of scientific, information-communication workers communities (González Barbone, 2002; Webster, 2002; Schiller, 1996; Harnad, 1995; Habermas, 1981); 3) the contracts between capitalistic governments, and corporations of the information recorded in documents such Thomson Reuters (and their indexes ISI, Journal Citations Reports (JCR), Web of Science, Web of Knowledge, etc.), as the the only standards to assess the worldwide science show evidence of the corruption between them (Herrán & Villena, 20012; Dilevko, 2009; González Barbone, 2002); it also shows their anti-scientific bias, since, for instance in Mexico between 1997-2006 the JCR, WOS, and WOK only indexed less than 15% of all the scientific information-communication (Torres Reyes, 2010), and these trends are similar throughout the world (Herrán & Villena, 2012); even the very same creator of the first scientific citation index (SCI) in 1995 has warned recently that to evaluate science only employing impact factor or JCR was limited and evend biased (Garfield, 2005); 4) while the largest percentage of worldwide innformation-communicationn (more than 85% only in Mexico, Torres Reyes, 2010) it is published thorugh Open Access journals, on the contrary, capitalistic corporations and States, and public institutions reviewers and financers of science (e.g. Mexican National Council of Science and Technology; CONACYT) refuse biasedly to include OA indexes (e.g. DOAJ, READLYC, SCIELO, PLOS, PUBMED, etc) as the major and best standards thus perpetuating the capitalistic corporation elitism. All the aforementioned in detriment of science, public universities and education, the public domain and sector, of the common good, and working society in general. Conclusions: It is projected that in 2600 there would be published 10 articles per second only in physics, that no one would have the time to read them (Hawking, 2002), just to mention only one discipline, thus there could be foreseen that if this trend of the capitalistic quantification, commodification, and evaluation of public information-communication against its quality, it goes against the scientific method per se, hence scientists in general and from sports in particular should promote quality by accessing-publishing in free of charge, communal, and democratic Open Access venues (e.g. Penrose, 2004, p. 1050; Harnad, 1994), because if this intervention of capitalistic Corporations-States within science and public universities and education would be still allowed, what is at stake is not only the privatization and commodification of public information-communication, but also of all the commons of public domain, e.g. water, air, democracy, education, libraries, etc. (Muela Meza, 2005; Marquand, 2004; Webster, 2002; Schiller, 1996); even the genomas of all human beings are at stake to be privatized and commodified (Orozco, 2008), thus then Orwellian eugenetics of fascism and totalitarianism will return, but more dangerous than that of the 20th Century
Facades of the Libreria di San Marco in Venice, The: An Interpretation of the Design Process
No full text"A new work in which I propose an interpretation of the design process Sansovino used to create the magnificent facades of the Libreria di San Marco in Venice, a masterpiece of Renaissance architecture." Sent to Marquand librarian by author Dec. 202
FPGA programming *New Date!
No full textThis lecture will cover the basics of FPGA programming. Explaining when FPGAs can solve performance issues in scientific computing, and illustrating tools and strategies to minimize development effort.
Requirements
Basic programming knowledge and general computer architecture and high-performance computing concepts.
Short Speaker Bio
Marco was a CERN openlab summer student in 2018 working on HCAL/ECAL energy reconstruction. He then moved to London and worked as a research software engineer at Maxeler Technologies. Now, Marco is a computing PhD student at Imperial College London and focusing on accelerating Monte Carlo simulations. He is recently working at Intel as a hardware modelling and simulation engineer. </p
Accelerating Large-Scale Graph Processing with FPGAs Lesson Learned and Future Directions
No full textProcessing graphs on a large scale presents a range of difficulties, including irregular memory access
patterns, device memory limitations, and the need for effective partitioning in distributed systems,
all of which can lead to performance problems on traditional architectures such as CPUs and
GPUs. To address these challenges, recent research emphasizes the use of Field-Programmable Gate
Arrays (FPGAs) within distributed frameworks, harnessing the power of FPGAs in a distributed
environment for accelerated graph processing. This paper examines the effectiveness of a multi-FPGA
distributed architecture in combination with a partitioning system to improve data locality and
reduce inter-partition communication. Utilizing Hadoop at a higher level, the framework maps the
graph to the hardware, efficiently distributing pre-processed data to FPGAs. The FPGA processing
engine, integrated into a cluster framework, optimizes data transfers, using offline partitioning for
large-scale graph distribution. A first evaluation of the framework is based on the popular PageRank
algorithm, which assigns a value to each node in a graph based on its importance. In the realm of
large-scale graphs, the single FPGA solution outperformed the GPU solution that were restricted
by memory capacity and surpassing CPU speedup by 26x compared to 12x. Moreover, when a
single FPGA device was limited due to the size of the graph, our performance model showed that a
distributed system with multiple FPGAs could increase performance by around 12x. This highlights
the effectiveness of our solution for handling large datasets that surpass on-chip memory restrictions
Art without an Author: Vasari’s Lives and Michelangelo’s Death
No full textMonografia sulla rappresentazione di Michelangelo nelle due edizioni delle Vite, sulla storia del libro e la questione della sua paternitàBook dedicated to the representation of Michelangelo in Vasari's Lives of the Artists, to the history of the book, and to the problem of its authorshi
Drug Utilization of Patients Admitted with Conditions Wholly Attributable to Alcohol in Friuli Venezia Giulia, Italy
No full text- …
