Previous | 
Home
Section: New Results
Process Placement with TreeMatch
We released TreeMatch version 1.0 in June. The new feature are: a stabilize API, optional integration of Scotch , extensive testing of all the features.
Application Domains
New Results
- Network Modeling
- Locality Aware Roofline Model
- Scalable Management of Platform Topologies
- New algorithm for I/O scheduling
- Topology-Aware Data Aggregation on Large-Scale Supercomputers
- Empirical Study of the Impact on Performance of Process Affinity and Metrics
- Automatic, Abstracted and Portable Topology-Aware Thread Placement
- Process Placement with TreeMatch
- Managing StarPU Communications with NewMadeleine
- New abstraction to manage hardware topologies in MPI applications
- Empirical Study of the Impact on Performance of Process Affinity and Metrics
- Gradient reconstruction in a legacy CFD application using task-based programming models
- Efficient multi-constraint graph partitioning algorithms
- Progress threads placement for MPI Non-Blocking Collectives
- Use of PaMPA on large-scale simulations
- Co-scheduling applications on cache-partitioned systems
- Dynamic memory-aware task-tree scheduling
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
Bibliography
Application Domains
New Results
- Network Modeling
- Locality Aware Roofline Model
- Scalable Management of Platform Topologies
- New algorithm for I/O scheduling
- Topology-Aware Data Aggregation on Large-Scale Supercomputers
- Empirical Study of the Impact on Performance of Process Affinity and Metrics
- Automatic, Abstracted and Portable Topology-Aware Thread Placement
- Process Placement with TreeMatch
- Managing StarPU Communications with NewMadeleine
- New abstraction to manage hardware topologies in MPI applications
- Empirical Study of the Impact on Performance of Process Affinity and Metrics
- Gradient reconstruction in a legacy CFD application using task-based programming models
- Efficient multi-constraint graph partitioning algorithms
- Progress threads placement for MPI Non-Blocking Collectives
- Use of PaMPA on large-scale simulations
- Co-scheduling applications on cache-partitioned systems
- Dynamic memory-aware task-tree scheduling
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
Bibliography