ROMA - 2016 - Annual activity report

ROMA

ROMA - 2016

Project-Team Roma

Members

Overall Objectives

Research Program

Algorithms for probabilistic environments
Platform-aware scheduling strategies
High-performance computing and linear algebra
Compilers, code optimization and high-level synthesis for FPGA

Application Domains

Applications of sparse direct solvers

Highlights of the Year

New Software and Platforms

MUMPS
DCC
PoCo
Aspic
Termite
Vaphor

New Results

A backward/forward recovery approach for the preconditioned conjugate gradient method
High performance parallel algorithms for the tucker decomposition of sparse tensors
Preconditioning techniques based on the Birkhoff–von Neumann decomposition
Parallel CP decomposition of sparse tensors using dimension trees
Scheduling series-parallel task graphs to minimize peak memory
Matrix symmetrization and sparse direct solvers
Robust Memory-Aware Mapping for Parallel Multifrontal Factorizations
Fast 3D frequency-domain full waveform inversion with a parallel Block Low-Rank multifrontal direct solver: application to OBC data from the North Sea
Matching-Based Allocation Strategies for Improving Data Locality of Map Tasks in MapReduce
Minimizing Rental Cost for Multiple Recipe Applications in the Cloud
Malleable task-graph scheduling with a practical speed-up model
Dynamic memory-aware task-tree scheduling
Optimal resilience patterns to cope with fail-stop and silent errors
Two-level checkpointing and partial verifications for linear task graphs
Resilient application co-scheduling with processor redistribution
A different re-execution speed can help
Coping with recall and precision of soft error detectors
Checkpointing strategies for scheduling computational workflows
Assessing General-Purpose Algorithms to Cope with Fail-Stop and Silent Errors
A failure detector for HPC platforms
Optimal multistage algorithm for adjoint computatio
Assessing the cost of redistribution followed by a computational kernel: Complexity and performance results
When Amdahl Meets Young/Daly
Computing the expected makespan of task graphs in the presence of silent errors
Toward an Optimal Online Checkpoint Solution under a Two-Level HPC Checkpoint Model
Cell morphing: from array programs to array-free Horn clauses
Symbolic Analyses of pointers
High-Level Synthesis of Pipelined FSM from Loop Nests
Estimation of Parallel Complexity with Rewriting Techniques

Bilateral Contracts and Grants with Industry

Bilateral Contracts with Industry
Technological Transfer: XtremLogic Start-Up

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: New Software and Platforms

PoCo

Polyhedral Compilation Library

Keywords: Polyhedral compilation - Automatic parallelization

Functional Description

PoCo (Polyhedral Compilation Library) is a compilation framework allowing to develop parallelizing compilers for regular programs. PoCo features many state-of-the-art polyhedral program analysis and a symbolic calculator on execution traces (represented as convex polyhedra). PoCo has been registered at the APP (“agence de protection des programmes”) and transferred to the XtremLogic start-up under an Inria licence.

Participant: Christophe Alias
Contact: Christophe Alias

Previous |

Home | Next next