FEniCS‘12: High-performance PDE frameworks for modern architectures - June 5-7, 2012
We are pleased to announce this year’s edition of the annual FEniCS workshop: FEniCS‘12: High-performance PDE frameworks for modern architectures June 5 – 7, 2012 at Simula Research Laboratory, Oslo, Norway. All FEniCS users, developers and other interested parties are invited to celebrate the recent release of FEniCS 1.0 and the FEniCS Book, and to discuss future directions for FEniCS development. A half-day FEniCS tutorial will be offered as part of the workshop. Participation in the workshop and tutorial is free of charge.
FEniCS‘12 at Simula Research Laboratory, June 5 – 7
We are pleased to announce this year’s edition of the annual FEniCS workshop:
FEniCS‘12: High-performance PDE frameworks for modern architectures June 5 – 7, 2012
at Simula Research Laboratory, Oslo, Norway.
All FEniCS users, developers and other interested parties are invited to celebrate the recent release of FEniCS 1.0 and the FEniCS Book, and to discuss future directions for FEniCS development. A half-day FEniCS tutorial will be offered as part of the workshop. Participation in the workshop and tutorial is free of charge.
We have open slots for contributed talks: if you are interested in presenting your recent work, please email a title and a short abstract to Marie Rognes (meg@simula.no) by May 1, 2012.
To register for the workshop, please contact Marie Rognes (meg@simula.no) by May 15, 2012.
Workshop abstract
Generic software frameworks play an increasingly important role in the numerical solution of partial differential equations. When new hardware architectures are introduced, these software frameworks must adapt to leverage the computational power of the new hardware and to match the performance of specialized hand-written code.
This workshop aims to bring together experts on generic software for partial differential equations and high performance computing on modern architectures. Topics include
User-friendly programming tools for many-core CPUs and GPUs
Automated code translation and generation targeting modern hardware
Hybrid computing using both CPUs and GPUs
Innovative PDE software techniques
Free/open-source software for numerical solution of PDEs
Invited speakers:
Scott Baden
Andre Brodtkorb
Carsten Burstedde
Filipe Cruz
Paul Kelly
Ridgway Scott
Andy Terrell
Garth Wells
Total number of participants: 46
Total number of guests outside of CBC: 21
Number of different nationalities represented: 7
Total number of speakers: 19
Total number of talks: 19
Program
Tuesday June 5
Location: "Storstua" @ Simula Research Laboratory
09:00 - 13:00 Registration (in Simula reception)
09:30 - 12:00 "FEniCS Tutorial" by Anders Logg
Session 1: Chair: Anders Logg
13:00 - 13:15 "Welcome" by Anders Logg
13:15 - 14:00 "Easy HPC" by Garth Wells
14:00 - 14:45 "Generating programs work better than transforming them, if you get the abstraction right" by Paul H. J. Kelly
14:45 - 15:00 Break
Session 2: Chair: Xing Cai
15:00 - 15:45 "Design of optimal Runge-Kutta methods" by David Ketcheson
15:45 - 16:15 Break
16:15 - 16:40 "Generating high-performance multi platform FEM solvers using Manycore Form Compiler and OP2" by Graham Markall
16:40 - 17:05 "An embedded language for vector operations via OpenCL" by Brian Brennan
17:05 - 17:30 "Algorithms for efficient compilation of complicated forms" by Martin S. Alnæs
Wednesday June 6
Location: "Storstua" @ Simula Research Laboratory
Session 3: Chair: Hans Petter Langtangen
10:00 - 10:45 "High-performance heterogeneous CPU-GPU computing on DEGIMA cluster" by Felipe Cruz
11:00 - 11:45 TBA by Andy Ray Terrel
12:00 - 13:00 Lunch
13:00 - 13:45 "Modular forest-of-octrees AMR: algorithms and interface" by Carsten Burstedde
13:45 - 14:30 "Scalable solution of non-linear time-dependent systems" by L. Ridgway Scott
14:30 - 14:45 Break
Session 4: Chair: Marie E. Rognes
14:45 - 15:10 "libadjoint: a new abstraction for developing adjoint models" by Simon Funke
15:10 - 15:35 "dolfin-adjoint: automating the adjoints of DOLFIN models" by David A. Ham
15:35 - 16:00 "GenFoo: a general Fokker-Planch solver with applications in fusion plasma physics" by Josef Höök
16:00 - 16:30 Break
16:30 - 16:55 "Nitsche’s method for fictitious domains and overlapping meshes: analysis and implementation" by Andre Massing
16:55 - 17:20 "The surprise talk" by Kent-Andre Mardal
19:00 - Dinner
Thursday June 7
Location: "Bakrommet" @ Simula Research Laboratory
Session 5: Chair: Xing Cai
09:30 - 10:15 "Compact stencils for shallow water equations on graphics processing units" by Andre R. Brodtkorb
10:15 - 11:00 "Computing at a million laptops per second" by Scott Baden
11:00 - 11:15 Break
11:15 - 12:00 "Auto-generating optimized CUDA for stencil computation" by Didem Unat
12:00 - 12:15 Closing remarks
Organizing committee:
Anders Logg
Marie Rognes
Xing Cai
Registration
By email to Marie Rognes (meg@simula.no) by May 15.
More information on the workshop can be found on the FEniCS webpage: http://fenicsproject.org/featured/2012/fenics_12_simula.html
Useful information for guests at CBC:
| What |
|
|---|---|
| When |
Jun 05, 2012 10:00 AM
to Jun 07, 2012 04:00 PM |
| Where | Simula |
| Contact Name | Marie Rognes |
| Attendees |
Amir Masoud Abdol Anders E. Johansen Anders Logg Andre Brodtkorb Andre Massing Andy Terrel Aron Ahmadia Bernardo L. deOliveira Brian Brennan Daniele Tartarini David Ham Didem Unat Eline Sundt Felipe Cruz Florian Rathgeber Fredrik Valdmanis Gabrial Balaban Garth Wells Glenn T. Lines Graham Marrkall Hans Petter Langtangen Ida Norderhaug Drosdal Joachim Berdal Haga Johan Hake Johannes Ring Josef Höök Jussi Koivumäki Kent-Andre Mardal Lyuda Vynnytska Marie E. Rognes Martin Alnæs Matteo Brunetti Mikael Mortensen Molly Maleckar Myles English Paul Kelly Rainer Nerlich Ridgway Scott Roberto Alessi Scott Baden Simon Funke Xing Cai |
| Add event to calendar |
vCal
iCal
|
