Organizing Committee
Abstract

As concurrency in scientific computing pushes beyond a million threads and performance of individual threads becomes less reliable for hardware-related reasons, attention of mathematicians, computer scientists, and supercomputer users and suppliers inevitably focuses on reducing communication and synchronization bottlenecks. Though convenient for succinctness, reproducibility, and stability, instruction ordering in contemporary codes is commonly overspecified. This workshop attempts to outline evolution of simulation codes from today's infra-petascale to the ultra-exascale and to encourage importation of ideas from other areas of computer science into numerical algorithms, new invention, and programming model generalization.

[Image courtesy of Hatem Ltaief] Directed Acyclic Graph for the execution of a symmetric generalized eigensolver for a dense 4x4 matrix

Confirmed Speakers & Participants

Workshop Schedule

Monday, January 9, 2012
TimeEventLocationMaterials
8:30 - 8:55Welcome Coffee & Registration11th Floor Collaborative Space 
8:55 - 9:00Welcome - Jill Pipher, Director, ICERM11th Floor Lecture Hall 
9:00 - 9:45Workshop Goals and Structure Meeting Logistics - David Keyes, KAUST11th Floor Lecture Hall 
9:45 - 10:45Lightning Talks (30 of two minutes each)11th Floor Lecture Hall 
10:45 - 11:00Coffee/Tea Break11th Floor Collaborative Space 
11:00 - 12:00Algorithmic Challenges of Exascale Computing - Katherine Yelick, University of California, Berkeley11th Floor Lecture Hall
12:00 - 2:00Working Lunch  
2:00 - 3:00Algorithm Design using Spectral Graph Theory - Gary Miller, CMU11th Floor Lecture Hall 
3:00 - 3:30Coffee/Tea Break11th Floor Collaborative Space 
3:30 - 4:00Lower bounds for communication in linear algebra - Grey Ballard, UC Berkeley11th Floor Lecture Hall 
4:00 - 4:30Scalable preconditioners for time-harmonic wave equations - Jack Poulson, University of Texas, Austin11th Floor Lecture Hall 
4:30 - 5:00An Integrative Model for Parallelism - Victor Eijkhout, University of Texas, Austin11th Floor Lecture Hall 
5:00 - 6:30Welcome Reception11th Floor Collaborative Space 
Tuesday, January 10, 2012
TimeEventLocationMaterials
9:00 - 10:00Multigrid at Extreme scales: Communication Reducing Data Models and Asynchronous Algorithms - Mark Adams, Columbia University11th Floor Lecture Hall 
10:00 - 10:30Coffee/Tea Break11th Floor Collaborative Space 
10:30 - 11:30Discussion and assignment of breakout groups11th Floor Lecture Hall 
11:30 - 12:00Initial Group MeetingsLocations will vary 
12:00 - 12:05Group Photo11th Floor Lecture Hall 
12:05 - 2:00Working Lunch  
2:00 - 3:00Sparse linear solvers on ever-wider machines - Jonathan Cohen, NVIDIA11th Floor Lecture Hall 
3:00 - 3:30Coffee/Tea Break11th Floor Collaborative Space 
3:30 - 4:00Principles of algorithm-architecture co-design - Richard Vuduc, Georgia Institute of Technology11th Floor Lecture Hall 
4:00 - 4:30Topical Talk V - Dinesh Kaushik, KAUST11th Floor Lecture Hall 
4:30 - 5:00Toward less synchronous composable multilevel methods for implicit multiphysics simulation - Jed Brown, Argonne National Laboratory11th Floor Lecture Hall
6:00 - 7:30Dinner - Directions are available at the 11th floor reception deskLocal 121 121 Washington Street Providence, RI 02903 
Wednesday, January 11, 2012
TimeEventLocationMaterials
9:00 - 10:00Finding the Happy Medium: Tradeoffs in communication, algorithms, architectures and programming models - Bill Gropp, University of Illinois at Urbana-Champain11th Floor Lecture Hall 
10:00 - 10:30Computational Challenges of Coupled Cluster Theory - Jeff Hammond, Argonne National Laboratory11th Floor Lecture Hall
10:30 - 10:55Coffee/Tea Break11th Floor Collaborative Space 
10:55 - 11:55Breakout GroupsLocations will vary 
11:55 - 1:40Working Lunch  
1:40 - 2:40Synchronization and Manycore Machines - John Owens, UC Davis11th Floor Lecture Hall 
2:40 - 3:10Making Dataflow Programming Ubiquitous for Scientific Computing - Hatem Ltaief, KAUST11th Floor Lecture Hall 
3:10 - 3:35Coffee/Tea Break11th Floor Collaborative Space 
3:35 - 4:35Preliminary plenary reports from Breakout Groups11th Floor Lecture Hall 
Thursday, January 12, 2012
TimeEventLocationMaterials
9:00 - 10:00Making Good Enough...Better: Addressing the Multiple Objectives of High-Performance Parallel Software with a Mixed Global-Local Worldview - John Gunnels, IBM11th Floor Lecture Hall 
10:00 - 10:30Automatic Selection of Compiler Optimizations Using Graph-Based Program Characterization - John Cavazos, University of Delaware11th Floor Lecture Hall 
10:30 - 11:00Coffee/Tea Break11th Floor Collaborative Space 
11:00 - 12:00Breakout GroupsLocations will vary 
12:00 - 2:00Break for Lunch  
2:00 - 3:00Hierarchical N-body algorithms: A pattern likely to lead at extreme scales - Lorena Barba, Boston University11th Floor Lecture Hall 
3:00 - 3:30Coffee/Tea Break11th Floor Collaborative Space 
3:30 - 4:00Group presentation I11th Floor Lecture Hall 
4:00 - 4:30Group presentation II11th Floor Lecture Hall 
4:30 - 5:00Group presentation III11th Floor Lecture Hall 
Friday, January 13, 2012
TimeEventLocationMaterials
9:00 - 10:00Towards a Science of Parallel Programming - Keshav Pingali, UT Austin11th Floor Lecture Hall 
10:00 - 10:30Coffee/Tea Break11th Floor Collaborative Space 
10:30 - 11:30Path forward from the point of view of mathematical challenges11th Floor Lecture Hall 
11:30 - 12:00Workshop Wrapup - David Keyes, KAUST11th Floor Lecture Hall 
12:00 - 12:00Adjournment  

Lecture Videos