HPC4EAS member participation at the CEDI 2013 (September 17-20) – Madrid.
Spanish Computer Conference (CEDI) aims to provide a framework for meeting professionals dedicated to research, development, innovation and university education in the field of computer engineering. The CEDI is structured as a multi-congress, consisting of a set of congresses, conferences or meetings that are being developed regularly.
- Metodología para predecir la escalabilidad de aplicaciones paralelas. J. Panadero, A. Wong, D. Rexachs & E. Luque. Proceedings XXIV JP2013. ISBN: 978-84-695-8330-2. p. 163-168.
- Performance & Scalability in Distributed Cluster-based Individual-oriented fish School Models by F. Borges, R. Solar, R. Suppi & E. Luque. Proceedings XXIV JP2013. ISBN: 978-84-695-8330-2. p. 402-407.
First Conference on Cloud Computing with the participation & tutorials of HPC4EAS group members. June 15-17, 2013. Computer Science School. UNLP.
- Cloud Security. Lic. Javier F. Diaz, LINTI – Universidad NAcional de La Plata. Argentina.
- Fault Detection and Correction in Cloud architectures. Dr. Dolores Rexach, Universitat Autònoma de Barcelona. Spain.
- Scalable Architectures for the Cloud. Intel. Argentina.
- Cloud Computing using Map-Reduce and Stream Computing platforms. Dr. Mauricio Marin, Universidad de Chile – Yahoo Research. Chile.
- Processing Scientific Applications in Cloud Computing Environments. Dr. García Garino, Universidad NAcional de Cuyo. Argentina.
- University perspective evaluating using cloud services. Dr. James A. Marsteller, Pittsburgh Supercomputing Center. USA.
The conference is complemented by courses taught in the postgraduate program of the UNLP:
- Cloud Computing. June, 10-14, 2013. Dr. Xoan Pardo. Universidade da Coruña. Spain.
- System Resource Management in High Performance Computing. June, 24-28, 2013. Dr. Remo Suppi. UAB. Spain.
- Fault Tolerant Computing in HPC. June, 24-28, 2013. Dr. Dolores Rexachs and Dr. Emilio Luque. UAB. Spain.
URL JCC13: http://jcc2013.info.unlp.edu.ar
PhD Thesis Defense: Sandra A.Méndez, Manuel Taboada González, Carlos H. Núñez Castillo, Marcela Castro León.
Sandra A.Méndez: Metodología para la evaluación de prestaciones del sistema Entrada/Salida en computadores de altas prestaciones. Date: July 11, 2013 at 9:30 am. Auditorium of the School of Engineering
Manuel Taboada González: Simulación de los Servicios de Urgencias Hospitalarias: una aproximación computacional desarrollada mediante técnicas de Modelado Orientadas al Individuo (Mol). Date: July 11, 2013 at 12:00pm. Auditorium of the School of Engineering
Carlos H. Núñez Castillo: Predictive and Distributed Routing Balancing for High Speed Interconnection Networks. Date: July 12, 2013 at 9:30 am. Auditorium of the Center for Computer Vision
Marcela Castro León: Tolerancia a fallos en la capa de sistema basada en la arquitectura RADIC. Date: July 11, 2013 at 17:00. Auditorium of the School of Engineering.
HPC4EAS member participation at the ICCS 2013 (Juny 5-7) – Barcelona.
The International Conference on Computational Science is an annual conference for researchers and scientists form different knowledge areas including (but not limited) to computer science, mathematics, life sciences, physics, chemistry, geology, etc is to say all applications areas who are pioneering in the utilization of computational methods in sciences as well as in arts and humanitarian fields, in order to discuss problems and solutions in the area, to identify new issues, and to shape future directions for research.
ICCS 2013 in Barcelona and is the thirteenth in this series of highly successful conferences http://www.iccs-meeting.org/iccs2013/
The first ICCS was in 2001 and has attracted increasingly higher quality and numbers of attendees and papers. The proceedings series have become a major intellectual resource for computational science researchers and serve to define and advance the state of the art of the field and the 30% high-quality papers for presentation at the conference and publication in the proceedings. These are published by Elsevier in their new Procedia Computer Science series and ICCS is an ERA 2010 A-ranked conference series.
The HPC4EAS Group members have presented the followings papers (in order of publication):
Predictive and Distributed Routing Balancing, an Application-Aware Approach
Pages 179-188
Carlos Núñez Castillo et al
Using an Agent-Based Simulation for Predicting the Effects of Patients Derivation Policies in Emergency Departments
Pages 641-650
Manel Taboada, et al.
Improving Communication Patterns for Distributed Cluster-Based Individual-Oriented Fish School Simulations
Pages 702-711
Roberto Solar, et al.
Evacuation Simulation Supporting High Level Behaviour-Based Agents
Pages 1495-1504
Pablo Cristian Tissera, et al.
A Tool for Selecting the Right Target Machine for Parallel Scientific Applications
Pages 1824-1833
Javier Panadero, et al.
Conference by Genaro Fernandes de Carvallo Costa, Associated Professor at Universidade Federal da Bahia, Brasil. Date: May 9, 2013. 12:00 hrs. Escola d’Enginyeria. UAB
Abstract: The daily social communication takes many forms of interaction. E-mails, tweets, text messages and social networks posts define the modern social lives of the individuals. And this information about the behavior can be collected, stored, and made available to researchers.
With access to billions of mails, tweets, and messages, is possible to build reliable models of social phenomena, like how the individuals choose their social partners o how a tweet spreads through a network.
There are many research projects with the objective of model human online behavior and find how it shapes social networks and how transform these patterns into a more general, abstract theory. Professor Costa present the working progress and results of the subject and remarkable ideas in this area.
Professor Costa is Associated Professor at UBBA and CEO at Convergence Works, PhD by UAB, and he has held positions of PostDoc at Federal University of Bahia, Postdoc at Universidade Federal da Bahia, Reseach Fellow at Universitat Autònoma de Barcelona, Manager at Telematic Company and CTO of Bahianet Internet e Serviços.
A proposal from Professor Luque, Mr. Taoboada and Dr. Valero was accepted for a grant award by IBM’s in the framework of Smarter Cities project.
IBM considers that a Smart cities know how to transform their systems and optimize the use of largely finite resources. To help drive efficiency and increase effectiveness, they leverage technology to make systems instrumented, interconnected and intelligent
- Instrumentation, or digitization, means that the workings of that system are turned into data points and the system is made measurable.
- Interconnection means that different parts of a core system can be joined and “speak” to each other, turning data into information.
- Intelligence refers to the ability to use the information created, model patterns of behavior or likely outcomes and translate them into real knowledge, allowing informed actions.
To help these objectives, the project led by IBM offers a set of services, tools and collaboration working groups that facilitate and will enable smart cities can become a reality (http://smartercitieschallenge.org/)
The proposal presented by this group include the following objectives:
- Research activities: Introduce HPC and Advanced Simulation in the Management and Optimization of Health Services.
Exploit the possibilities opened by this research field developing, in cooperation with IBM, powerful tools such as Decision Support Systems (DSS), for smarter management in the health services area. - Teaching activities: Introducing Advanced Simulation and Health Services management in the Computer Science and Services degree.
Increasing the IBM cooperation, as an educational partner, in the areas of HPC and services for the student of the CS&S degree.
By Professor Alba Cristina M. A. de Melo. Department of Computer Science. University of Brasilia.
Federated Cloud Computing: Power-Aware Solutions and Applications
Abstract: Cloud computing is a recent paradigm for provision of computing infrastructure, platform and/or software in the Internet, which aims to reduce costs. One important factor in the operational cost of cloud data centers comes from the power consumed by the computational resources. Many studies have been conducted to provide power reduction but most of the proposed techniques incur in a great slowdown of the applications. In this talk, I will first address aspects of Cloud Computing and Green Computing. Then, I will present our server consolidation strategy for Federated Cloud systems. I will also show you some simulation results that show that our collaborative strategy is able to reduce more than 50% of energy consumption, while still meeting the Quality-of-Service requirements. After that, I will present one strategy to run a huge Bioinformatics application on Federated Public Clouds. Finally, I will present the conclusions and future directions.
Professor Alba Cristina Magalhaes Alves de Melo received the Ph.D. degree in Computer Science from the Institut National Polytechnique de Grenoble, France, in 1996, the Ms.C. degree in Computer Science from UFRGS, Brazil, in 1991 and the BS degree in Computer Science from UnB, Brazil, in 1986. She is currently an Associate Professor at the University of Brasilia, Brazil, and CNPq Fellow Researcher level 1D. Her current research interests include cloud, grid, cluster and P2P computing, bioinformatics and application-specific accelerators. She has advised 18 graduate students in these research areas and published several papers in prestigious international journals and conferences. She is a Senior Member of the IEEE Society and the IEEE Computer Society.
By Professor Armando De Giusti. Faculty of Informatics. Universidad Nacional de La Plata, Argentina. Director of the Institute for Research in Computer LIDI. By Professor Marcela Printista. Computer Science Departament. Universidad Nacional de San Luis.
Professor De Giusti will present the progress of research lines developed in the III-LIDI with particular focus on the areas: HPC, Cluster & Cloud Computing, Distributed & Parallel Systems and Parallel Algorithms.
Professor Marcela Prinstina is staff member of the R&D Laboratory in Computational Intelligence (LIDIC) and perfoms research in Evacuation Simulations using Cellular Automata. Computer simulations using Cellular Automata (CA) have been applied with considerable success in different scientific areas, such as chemistry, biochemistry, economy, physics, etc. Professor Prinstina will present the work developed using CA in order to specify and implement a simulation model that allows to investigate behavioral dynamics for pedestrians in an emergency evacuation.
Conference by Barton P. Miller. Professor of Computer Sciences Department at University of Wisconsin.
Research in Cyber Security
The security team led by Prof. Barton Miller at the University Wisconsin is working on several areas of research that we describe during talk. The first area is the analysis, monitoring, and control of malicious programs. The research is based on a hybrid static/dynamic technique that monitors known code in a binary, and then discovers new code as it is decoded, unpacked, and modified, guaranteeing that we can monitor and control the code before it is executed. This research is being incorporated in the Dyninst binary instrumentation tool suite. The second area is the use of advanced machine learning techniques to expose the provinance of binary programs. These techniques can report the source language(s) in which the program was written, the compiler used and the optimization level. In addition, we can also identify the author of the program solely from the binary code. The last area of research is our joint work with the Autonomous University of Barcelona on the in-depth vulnerability assessment of middleware and services. The work at Wisconsin includes a technique called self-propelled instrumentation, that can be injecting into a running program and trace its behavior, and propagate this tracing in other processes, even on other hosts.
The Workshop will be held at the Escola d’Enginyeria, Universitat Autònoma de Barcelona between April 8 and April 12, 2013 . The main objectives of this worshop is to present the state of the art and the researchs advances in the CAPITA Project.12/4/2013. NEW! : Gallery View
The CAPITA project integrates a series of lines of research, interrelated and developed in the context of high performance computing (HPC) such as:
1. Performance and Efficiency in the use of HPC resources
2. User Availability of HPC Resources
3. Design and optimization of HPC systems, for “workloads” specific (application-specific domains)
4. Social projection (impact) applications:
(Consult Agenda for schedule).
Location:
Conference by Dr Mauricio Marin, Yahoo! Labs Santiago Researcher, Professor at Universidad de Santiago, Chile
Abstract: Large scale Web search engines are complex and highly optimized systems devised to operate on dedicated clusters of processors. Any, even a small, gain in performance is beneficial to economical operation given the large amount of hardware resources deployed in the respective data centers. Performance is fully dependent on users behavior which is featured by unpredictable and drastic variations in trending topics and arrival rate intensity. In this context, discrete event simulation is a powerful tool either to predict performance of new optimizations introduced in search engine components or to evaluate different scenarios under which alternative component configurations are able to process demanding workloads. These simulators must be fast, memory efficient and parallel to cope with the execution of millions of events in small running time on few processors. We propose achieving this objective at the expense of performing approximate optimistic parallel discrete-event simulation.
Raising to the clouds by Eduardo Argollo (Research Scientist HP Exascale Computing Lab, Sant Cugat del Valles)
Eduardo Argollo is Research Scientist at HP Labs since 2007 (Computer Science MS degree in 1997, Catholic University of Salvador, in Brazil, PhD in Computer Science from the University Autonoma of Barcelona, in Spain)
Eduardo has more than 10 years of experience in the ERP systems, databases and data warehousing development. Prior to joining HP, he was worked in methodologies to efficiently attain high-performance from a collection of internet-connected geographically distributed multi-clusters and in the CoreGRID Network of Excellence in Passau, Germany. In this work was involved in the automatic parallelization of applications using higher-order Grid components.
His current research interests include system-level simulation, virtualization and parallel computing middleware.
Researchers of the our group (HPC4EAS) in collaboration with the team at the Emergency Services Unit at Hospital de Sabadell (Parc Taulí Healthcare Corporation), have developed an advanced computer simulator to help in decision-making processes (DSS, or decision support system) which could aid emergency service units in their operations management.
The model was designed based on real data provided by the Parc Taulí Healthcare Corporation, using modelling and simulation techniques adapted to each individual, and which require the application of high performance computing. The system analyses the reaction of the emergency unit when faced with different scenarios and optimises the resources available.
Researchers defined different types of patients according to their emergency level, and doctors, nursing teams, and admissions staff according to different levels of experience. This permitted studying the duration of processes such as the triage (when the emergency level is determined), the number and type of patients arriving at each moment, the waiting period for each stage or phase of the service, costs associated with each process, the amount of staff needed to determine a type of assistance and, in general, all other quantifiable variables. The system not only helps to make decisions in real time, it also can help by making forecasts and improving the functioning of the service.
PhD Thesis Defense: MªMar López, Date: Sept. 13, 11:00 hrs. 2012. Planificación de DAGs en entornos oportunísticos.
Las aplicaciones tipo workflow se caracterizan por tener un elevado tiempo de cómputo y una elevada transferencia de datos. Como consecuencia, el tiempo de ejecución o makespan de un workflow es elevado. Con el propósito de reducir el makespan del workflow, las tareas se ejecutan en diferentes máquinas interconectadas a través de una red. Asignar correctamente las tareas del DAG a las máquinas disponibles del entorno de ejecución mejora el makespan. El encargado de realizar la asignación de las tareas del workflow a las máquinas es el planificador.
El problema de un planificador estático es que no tiene en cuenta los cambios ocurridos en el entorno de ejecución durante la ejecución del DAG.
La solución a este problema ha sido el desarrollo de un nuevo planificador dinámico.
El planificador dinámico mejora el makespan del DAG debido a que considera los cambios ocurridos en el entorno de ejecución durante la ejecución del workflow, pero como contrapartida, genera overhead producido a consecuencia de reaccionar ante los cambios detectados. El objetivo de este trabajo es proporcionar estrategias que reducen el overhead del planificador dinámico, sin afectar al makespan del DAG. Para reducir el overhead, el algoritmo reacciona ante los cambios detectados durante la ejecución del DAG únicamente si anticipa que su makespan mejora.
La política dinámica desarrollada ha sido evaluada a través de ejecuciones simuladas y ejecuciones realizadas en un entorno oportunístico real. En la experimentación simulada se ha mejorado el makespan entre 5% y 30%, y en la experimentación real la mejora del makespan ha sido entre 5% y 15%. En lo que respecta al overhead, éste se ha reducido como mínimo un 20% respecto a otras políticas de planificación dinámicas.
Journal of Computational Science. Best Paper Award. 2012
For the second year (consecutive) HPC4EAS research group receives the Best Paper Award for the article Simulation Optimization for Helathcare Emergency Departaments. Authors: Cabrera E., Taboada M., Iglesias ML., Epelde F., Luque E.
In 2011 the group was awarded with the Outstanding Paper Award at the ICCS 2011 (Singapore) for the article High Performance Distributed Cluster-Based Individual Oriented Fish Schools Simulation. Authors: Solar R., Suppi R., Luque E.