Program
Time |
Event |
|
09:30 - 09:45
|
Welcoming talk (Amphitheatre Leyteire) - Organizers |
|
09:45 - 10:30
|
Plenary talk (Amphitheatre Leyteire) - Olivier Beaumont |
|
09:45 - 10:30 |
› HPC-BigData Convergence: What to do when scientific data becomes too big? - Franck Cappello, Argonne National Laboratory [Lemont] |
|
10:30 - 11:00
|
Break (Hall Leyteire) |
|
11:00 - 12:40
|
Numerical Algorithm (Amphitheatre Leyteire) - Grégoire Pichon |
|
11:00 - 11:25 |
› A 2D Task-based Fan-both Factorization Algorithm for Sparse Symmetric Matrices - Mathias Jacquelin, Lawrence Berkeley National Laboratory |
|
11:25 - 11:50 |
› Improving nested task-parallelism in the LU Factorization of H-Matrices - Rocío Carratalá-Sáez, Universitat Jaume I |
|
11:50 - 12:15 |
› Extending the STF model to design a hierarchical direct solver for distributed memory machines with manycore nodes - Cedric Augonnet, CEA/DAM |
|
12:15 - 12:40 |
› Wine or Brandy? Block low rank and block separable matrices - Cleve Ashcraft, Livermore Software Technology Corporation |
|
12:40 - 14:00
|
Lunch (Hall Leyteire) |
|
14:00 - 15:15
|
Scheduling (Amphitheatre Leyteire) |
|
14:00 - 14:25 |
› Independent tasks on 2 resources with co-scheduling effects - Lionel Eyraud-Dubois, Inria Bordeaux - Sud-Ouest |
|
14:25 - 14:50 |
› Optimal Memory-aware Backpropagation of Deep Join Networks - Julien Herrmann, Inria Bordeaux |
|
14:50 - 15:15 |
› Computing latency of a real-time system modeled by a Synchronous Dataflow Graph - Alix Munier Kordon, Laboratoire dÍnformatique de Paris 6 |
|
15:15 - 15:40
|
Discussions - In Amphitheatre Leyteyre and rooms nearby |
|
15:40 - 16:10
|
Coffee break (Hall Leyteire) |
|
16:10 - 17:50
|
Performance Evaluation (Amphitheatre Leyteire) - Abdou Guermouche |
|
16:10 - 16:35 |
› A Comparison of Random Task Graph Generation Methods for Scheduling Problems - Louis-Claude Canon, Département d'Informatique des Systèmes Complexes (FEMTO-ST/DISC) |
|
16:35 - 17:00 |
› Faithful and Efficient Simulation of High Performance Linpack - Arnaud Legrand, Laboratoire d'Informatique de Grenoble |
|
17:00 - 17:25 |
› Is your scheduling good? How would you know? - Anthony Danalis, University of Tennesee |
|
17:25 - 17:50 |
› An overview of uncore frequency impact on performance and power consumption - Amina Guermouche, Telecom SudParis |
|
19:30 - 22:00
|
Dinner - St-Once -
The first dinner party will take place at the Domaine du Haut-Carré outside if the weather is nice, and in the nearby room Badiane otherwise. Food, water and juices will be served, but no alcoholic beverages are provided. It is a bring your own wine party to discover wines from all over the world. |
|
Time |
Event |
|
08:30 - 09:00
|
Breakfast (Hall Leyteire) |
|
09:00 - 10:15
|
Scheduling (Amphitheatre Leyteire) - Taieb Znati |
|
09:00 - 09:25 |
› Scheduling at the edge - Clement Mommessin, Inria |
|
09:25 - 09:50 |
› Adaptive Parallelism for Coupled, Multithreaded Message-Passing Programs - Dorian Arnold, Emory University [Atlanta, GA] |
|
09:50 - 10:15 |
› Experiences in the design of a malleable and I/O-aware scheduler - David Singh, Carlos III University of Madrid |
|
10:15 - 10:45
|
Break (Hall Leyteire) |
|
10:45 - 12:00
|
Resilience (Amphitheatre Leyteire) - Frédéric Vivien |
|
10:45 - 11:10 |
› Partial Redundancy and In-memory checkpoints Under Heterogeneous Failure Likelihoods - Rami Melhem, The University of Pittsburgh |
|
11:10 - 11:35 |
› A nice little scheduling problem - Yves Robert, Laboratoire de lÍnformatique du Parallélisme, Innovative Computing Laboratory [Knoxville] |
|
11:35 - 12:00 |
› Fault-Tolerance Strategies for HPC Platforms - Thomas Herault, University of Tennessee |
|
12:00 - 13:30
|
Lunch (Hall Leyteire) |
|
13:30 - 14:45
|
Job Scheduling (Amphitheatre Leyteire) |
|
13:30 - 13:55 |
› Adaptive scheduling of HPC applications using malleability and dynamic migration. - ALBERTO CASCAJO, University Carlos III of Madrid - Jesus Carretero, University Carlos III of Madrid |
|
13:55 - 14:20 |
› Reservation Strategies for Stochastic Jobs |
|
14:20 - 14:45 |
› Strategies for deploying stochastic HPC applications on large-scale systems - Ana Gainaru, Vanderbilt University [Nashville] |
|
14:45 - 16:00
|
Discussions - In Amphitheatre Leyteire and nearby rooms |
|
16:00 - 16:30
|
Coffee break (Hall Leyteire) |
|
16:30 - 17:45
|
Graphs (Amphitheatre Leyteire) - Esmond Ng |
|
16:30 - 16:55 |
› Parallel Algorithms for Dynamic Networks - Sanjukta Bhowmick, University of North Texas |
|
16:55 - 17:20 |
› Is Acyclic Directed Graph Partitioning Effective for Locality-Aware Scheduling? - Ümit Çatalyürek, Georgia Institute of Technology |
|
17:20 - 17:45 |
› Graph coloring and sparse matrix multiplication: a hammer and an incompatible nail - Aydin Buluc, Lawrence Berkeley National Laboratory |
|
19:30 - 22:00
|
Dinner - Chez Frida -
The dinner will take place at Chez Frida, 27-29 rue Buhan
Dinner will start at 8PM. |
|
Time |
Event |
|
08:30 - 09:00
|
Breakfast (Hall Leyteire) |
|
09:00 - 10:40
|
Numerical Algorithm (Amphitheatre Leyteire) - Théo Mary |
|
09:00 - 09:25 |
› Sparse tensor ordering - Bora Uçar, ROMA |
|
09:25 - 09:50 |
› Optimal tensor-matrix and -vector contraction strategies for computing dense tensor decompositions - Oguz Kaya, Laboratoire de Recherche en Informatique |
|
09:50 - 10:15 |
› Parallel QR factorization of block-tridiagonal matrices from Generative Models - Alfredo Buttari, Institut de recherche en informatique de Toulouse |
|
10:15 - 10:40 |
› This is yet another talk about scheduling tricks for QR factorization - Julien Langou, Department of Computer Science and Engineering [Denver] |
|
10:40 - 11:00
|
Break (Hall Leyteire) |
|
11:00 - 12:40
|
Job Scheduling (Amphitheatre Leyteire) - Kaya Oguz |
|
11:00 - 11:25 |
› Efficient Job Scheduling for Clusters with Shared Tiered Storage - Hamid Mohammadi Fard, Technische Universität Darmstadt |
|
11:25 - 11:50 |
› Power-aware scheduling for Large-scale Supercomputers - Tapasya Patki, Lawrence Livermore National Laboratory |
|
11:50 - 12:15 |
› Optimal Vaccination Schedules in Large SIR Systems from a Central and an Agent's Perspective. - Bruno Gaujal, Inria Grenoble Alpes |
|
12:15 - 12:40 |
› Optimal Online Load Maximization with Commitment - Uwe Schwiegelshohn, TU Dortund University |
|
12:40 - 14:00
|
Lunch (Hall Leyteire) |
|
14:00 - 15:40
|
DAG Scheduling (Amphitheatre Leyteire) - Mathieu Faverge |
|
14:00 - 14:25 |
› An EPTAS for scheduling fork-join task graphs with communication delays - Oliver Sinnen, University of Auckland |
|
14:25 - 14:50 |
› Scheduling Task Graphs on Bounded Shared-Memory Platforms - Loris Marchal, Laboratoire de lÍnformatique du Parallélisme |
|
14:50 - 15:15 |
› Task graph execution managed by a runtime: an opportunity for theoreticians - Samuel Thibault, Université de Bordeaux |
|
15:15 - 15:40 |
› SPAGHETtI: Scheduling/Placement Approach for task-Graphs on HETerogeneous archItecture - Emmanuel Jeannot, Inria Bordeaux - Sud-Ouest |
|
15:40 - 16:10
|
Coffee break (Hall Leyteire) |
|
16:10 - 17:50
|
Discussions - In Amphitheatre Leyteire and nearby rooms |
|
|