RUNTIME - 2011 - Annual activity report

RUNTIME

RUNTIME - 2011

Project Team Runtime

Members

Overall Objectives

Designing Efficient Runtime Systems
Highlights

Scientific Foundations

Runtime Systems Evolution

Application Domains

Application Domains

Software

Common Communication Interface
Hardware Locality
KNem
Marcel
ForestGOMP
Open-MX
StarPU
NewMadeleine
PadicoTM
MAQAO
QIRAL
TreeMatch

New Results

High-Performance Intra-node Collective Operations
I/O-Affinity-aware MPI Communications
High-Performance Point-to-Point Communications
Improve code-coupling performance in the SALOME platform
Hardware topology-aware MPI applications
Mastering Heterogeneous Platforms
Development of a flexible heterogeneous system-on-chip platform using a mix of programmable processing elements and hardware accelerators
Sparse GMRES on heterogeneous platforms in oil extraction simulation
Programming models for heterogeneous platforms
Parallel Concha
Scientific Application Analysis and Experiments
Virtualization of GPUs for OpenCL
Automatically Adaptating Task Grain for Hybrid Architectures
Performance modeling for power consumption reduction on the SCC
Modeling cache coherence protocol overhead
Memory Performance Analysis and Tool for OpenMP codes
Data-layout Optimization for Stencil codes on multi-cores and GPUs

Contracts and Grants with Industry

Contracts with Industry

Partnerships and Cooperations

National Initiatives
European Initiatives
International Initiatives

Dissemination

Animation of the scientific community
Seminars and invited talks
Teaching
Diffusion of the scientific culture

Bibliography

Major publications
Publications of the year
References in notes

Home | Next next

2011 Team Activity Report

RUNTIME

Efficient runtime systems for parallel architectures

Research centre: Bordeaux - Sud-Ouest

In partnership with: CNRS, Ecole nationale supérieure d'électronique, informatique et radiocommunications de Bordeaux, Université de Bordeaux

In collaboration with: Laboratoire Bordelais de Recherche en Informatique (LaBRI)

Field: Networks, Systems and Services, Distributed Computing

Theme: Distributed and High Performance Computing

Keywords: Grid Computing, High Performance Computing, Scheduling, Grid'5000, High Performance Communication, Multithreading