A5059115581- Distributed and Parallel Computing Systems
- Parallel Computing and Optimization Techniques
- Scientific Computing and Data Management
- Advanced Data Storage Technologies
- Cloud Computing and Resource Management
- Service-Oriented Architecture and Web Services
- Embedded Systems Design Techniques
- Interconnection Networks and Systems
- Distributed systems and fault tolerance
- Simulation Techniques and Applications
- Modular Robots and Swarm Intelligence
- Business Process Modeling and Analysis
- Petri Nets in System Modeling
- Electromagnetic Scattering and Analysis
- Network Time Synchronization Technologies
- Advanced Optical Network Technologies
- Satellite Communication Systems
- Advanced Software Engineering Methodologies
- Real-Time Systems Scheduling
- Software System Performance and Reliability
- Network Traffic and Congestion Control
- Bayesian Modeling and Causal Inference
- Information Technology and Learning
Centre National de la Recherche Scientifique
1996-2024
Laboratoire Bordelais de Recherche en Informatique
2005-2024
Université de Bordeaux
2009-2024
Institut Polytechnique de Bordeaux
2009-2023
Centre de Recherche Inria Bordeaux - Sud-Ouest
2023
Alternative Opportunities
2006-2014
Institut national de recherche en informatique et en automatique
1995-2012
Imperial College London
2001-2005
Université Côte d'Azur
1996-2002
The increasing numbers of cores, shared caches and memory nodes within machines introduces a complex hardware topology. High-performance computing applications now have to carefully adapt their placement behavior according the underlying hierarchy resources software affinities. We introduce Hardware Locality (hwloc) which gathers information about processors, caches, more, exposes it runtime systems in abstracted portable hierarchical manner. hwloc may significantly help performance by...
The emergence of accelerators as standard computing resources on supercomputers and the subsequent architectural complexity increase revived need for high-level parallel programming paradigms.Sequential task-based model has been shown to efficiently meet this challenge a single multicore node possibly enhanced with accelerators, which motivated its support in OpenMP 4.0 standard.In paper, we show that paradigm can also be employed achieve high performance modern composed multiple such nodes,...
The move towards Service Grids, where services are composed to meet the requirements of a user community within constraints specified by resource provider, present many challenges service provision and description. To support our research activities in autonomous composition form Semantic Grid we describe adoption ICENI web enable interoperability with recently proposed Open Services Architecture.
As the number of cores per machine increases, memory architectures are being redesigned to avoid bus contention and sustain higher throughput needs. The emergence Non-Uniform Memory Access (NUMA) constraints has caused affinities between threads buffers become an important decision criterion for schedulers. migration dynamically enables joint distribution work data across but requires high-performance transfers as well a convenient programming interface. We present improvements LINUX...
The move towards Service Grids, where services are composed to meet the requirements of a user community within constraints specified by resource provider, present many challenges service provision and description. To support our research activities in autonomous composition form Semantic Grid we describe adoption ICENI web enable interoperability with recently proposed Open Services Architecture.
Effective exploitation of computational grids can only be achieved when applications are fully integrated with the grid middleware and underlying resources. Fundamental to this is information. Information about structure behavior application, capability networking resources, availability access these resources by an individual, a group or organization. This paper describes environment that open, extensible platform independent. We match high-level application specification, defined as...
Communication libraries have dramatically made progress over the fifteen years, pushed by success of cluster architectures as preferred platform for high performance distributed computing. However, many potential optimizations are left unexplored in process mapping application communication requests onto low level network commands. The fundamental cause this situation is that design subsystems mostly focused on reducing latency shortening critical path. In paper, we present a new scheduling...
The advancement of modelling and simulation within complex scientific applications is currently constrained by the rate at which knowledge can be extracted from data produced. As Grid computing evolves, new means increasing efficiency analysis are being explored. RealityGrid aims to enable more efficient use resources condensed matter, materials biological science communities. Imperial College e-Science Networked Infrastructure (ICENI) middleware provides an end-to-end pipeline that...
In this paper, we present two new methods to simulate Petri Nets: a data parallel simulation and distributed simulation. Both simulations use an equational representation of the net in so called (min, +) algebra. The is based on matrix these equations, decomposition Net into marked graph components.
The goal of Matrices Over Runtime Systems at Exascale (MORSE) project is to design dense and sparse linear algebra methods that achieve the fastest possible time an accurate solution on large-scale multicore systems with GPU accelerators, using all processing power future high end can make available. In this poster, we propose a framework for describing algorithms level abstraction delegating actual execution runtime system in order software whose performance portable accross architectures....