PADAL19‎ > ‎


Monday September 9th
  9:00 Registration and Welcome
  9.30 Keynote by Samuel Thibault - Data Locality in Task Scheduling, a Runtime System Perspective
10:30 Break
11:00 Session 1 (3 talks)
12:00 Lunch Break
13:00 Session 1 - cont. (3 talks)
14:00 Panel for Session 1
15:00 Break
15:30 Session 2
16:50 Panel for Session 2
17:50 Adjourn

Tuesday September 10th
  9.00 Session 3 (4 talks)
10:20 Break
10:50 Session 3 - cont. (1 talk)
11:10 Panel for Session 3
12:10 Lunch Break
13:10 Session 4
14:30 Break
15:00 Panel for Session 4
16:00 Adjourn

Wednesday September 11th
  9.10 Session 5 (4 talks)
10:30 Break
11:00 Session 5 - cont. (1 talk)
11:20 Panel for Session 5
12:20 Lunch Break

Presentations are available here.

Session 1 Programming Models and Runtime Systems (Abstracts)
  1. John Shalf, Data Locality for Extreme Heterogeneous Systems
  2. François Tessier, MAESTRO: Middleware for Memory and Data-Awareness in Workflows
  3. Ron Brightwell, Thoughts on Autonomous Resource Management for HPC
  4. Karim Djemame, Multi-Threading Effective Task Scheduling for Heterogeneous Computing
  5. Mustafa Abduljabbar, Runtime-assisted locality abstraction using virtual topologies: status and roadmap
  6. Christian Terboven, Programming for Affinity

Session 2 DSL/Compilers (Abstracts)
  1. Kentaro Sano, FPGA Cluster as Custom Computing Engine for Supercomputers
  2. Alexandros Nikolaos Ziogas, A Data-Centric Approach to High Performance Computing Applications
  3. Mohamed Wahib, A Versatile Software Systolic Execution Model for GPU Memory-Bound Kernels
  4. Paul H J Kelly, Getting the mesh abstraction right: Delivering and generalising domain-specific optimisations for mesh-based computations

Session 3 Performance Models and Tools (Abstracts)
  1. Anshu Dubey, Cost Model for Application Configuration with Heterogeneity
  2. Aleksandar Ilic, Cache-Aware Roofline Model: Uncovering micro-architecture upper-bounds
  3. Emmanuel Jeannot, TopoMatch: a generic tool for topology mapping
  4. Vitus J Leung, Level-Spread: A New Job Allocation Policy for Dragonfly Networks
  5. Didem Unat, Communication monitoring for threads

Session 4 Apps and Core Kernels (Abstracts)
  1. Hatem Ltaief, Low-Rank Cholesky Factorization Toward Climate and Weather Prediction Applications
  2. Xing Cai, Unstructured computational meshes and data locality
  3. Mauro Bianco, Halo-Update Communication Layer for Hybrid Computing
  4. Naoya Maruyama, Constructing a Streaming SpMV Kernel for FPGAs

Session 5 Memory Management (Abstracts)
  1. Mercier Guillaume, Locality and Hardware Topology Management in MPI
  2. Brice Goglin, Exposing Heterogeneous Memory Characteristics to HPC Applications
  3. Balazs Gerofi, Towards Intelligent Management of Heterogeneous Memory: A Reinforcement Learning Approach
  4. Julien JAEGER, Profile-guided scope-based data allocation method for heterogeneous memory architecture
  5. Bert Wesarg, Memory Layout Abstraction in Modern C++ for HPC