TADAAM - 2019 - Annual activity report

TADAAM

TADAAM - 2019

Project-Team Tadaam

Team, Visitors, External Collaborators

Overall Objectives

Research Program

Need for System-Scale Optimization
Scientific Challenges and Research Issues

Application Domains

Mesh-based applications

Highlights of the Year

New Software and Platforms

Hsplit
hwloc
NetLoc
NewMadeleine
PaMPA
TopoMatch
SCOTCH
disk-revolve

New Results

Management of heterogeneous and non-volatile memories in HPC
Modeling and Visualizing Many-core HPC Platforms
Co-scheduling HPC workloads on cache-partitioned CMP platforms
Modeling High-throughput Applications for in situ Analytics
Modeling Non-Uniform Memory Access and Heterogeneous Memories on Large Compute Nodes with the Cache-Aware Roofline Model
Statistical Learning for Task and Data Placement in NUMA Architecture
On-the-fly scheduling vs. reservation-based scheduling for unpredictable workflows
Scheduling strategies for stochastic jobs
Online Prediction of Network Utilization
An Introspection Monitoring Library
Tag matching in constant time
Dynamic broadcasts in StarPU/NewMadeleine
Task based asynchronous MPI collectives optimisation
Dynamic placement of progress thread for overlapping MPI non-blocking collectives on manycore processor
Dynamic placement of Hybrid MPI +X coupled applications
Scheduling on Two Unbounded Resources with Communication Costs
H-Revolve: A Framework for Adjoint Computation on Synchrone Hierarchical Platforms
Sizing and Partitioning Strategies for Burst-Buffers to Reduce IO Contention
Optimal Memory-aware Backpropagation of Deep Join Networks
Optimal checkpointing for heterogeneous chains: how to train deep neural networks with limited memory
I/O scheduling strategy for HPC applications
A New Framework for Evaluating Straggler Detection Mechanisms in MapReduce
Clarification of the MPI semantics
Adaptive Request Scheduling for the I/O Forwarding Layer using Reinforcement Learning

Bilateral Contracts and Grants with Industry

Bilateral Grants with Industry

Partnerships and Cooperations

Regional Initiatives
National Initiatives
European Initiatives
International Initiatives
International Research Visitors

Dissemination

Promoting Scientific Activities
Teaching - Supervision - Juries
Popularization

Bibliography

Publications of the year
References in notes

Previous |

Home | Next next

Section: Partnerships and Cooperations

Regional Initiatives

CRA HPC Scalable Ecosystem, 2018-2021

2018 - 2021 (36 months)
Coordinator: Emmanuel Agullo
Other partners: INRA, Institut Pprime, UPPA, Airbus, CEA, CATIE
Abstract: The goal is to design a unified runtime-system for numerical simulation at large-scale and with a large amount of data. We aim at contributing significantly to the convergence between HPC and BigData. TADaaM is involved in scheduling data access and managing communication efficiently on large-scale system.

Previous |

Home | Next next