2011 Agenda

9/7/2011(Wednesday)Session ChairSpeaker
8:30 Workshop Introduction and Opening Remarks Chris Kerr  
9:00 Session 1- Challenges of using hybrid multi-core systems Rich Loft  
  Towards Exascale - Developments in Hardware and Software   Mike Ashworth
  The Keeneland Project   Jeff Vetter
  Successes and Challenges using GPUs for Weather and Climate Models   Mark Govett
10:00 Discussion period 1    
10:30 Morning Break    
10:45 Session 2 - Applications and Frameworks Bill Putman  
  Opportunity of improving WRF Performance thru heterogeneous computing   Charles Finan
  Experience using FORTRAN GPU Compilers with the NIM Tom Henderson
  Challenges of a Climate Modeling Infrastructure as a Research Tool on Exascale Platforms   Jeff Durachta
11:45 Discussion period 2    
12:15 Lunch    
1:15 Session 3 - Programming Approaches for Physics Packages Ilene Carpenter  
  Early Performance Analysis of the CMI code with LFOICE microphysics on the Power 7   Mark Straka
  GPU Acceleration of the RRTM in WRF using CUDA FORTRAN   Greg Ruetsch
  Lessons Learned adapting GEOS-5 GCM Physics to CUDA FORTRAN   Matt Thompson
  Accelerated Cloud Resolving Model in Hybrid CPU-GPU Clusters   Jose Garcia
2:35 Discussion period 3    
3:05 Afternoon Break    
3:15 Session 4 - Programming Tools Chris Kerr  
  Programming Tools: Experiences using programming , debugging and performance tools   Richard Gerber
  Allinea Tools   Dave Maples
  Status and Directions of NVIDIA GPUs for Earth Modeling Systems Stan Posey
  Facing the Debugging Challenges of Heterogeneous and Multi-core Programming   Ed Hinkel
4:35 Discussion period 4    
5:10 Workshop First Day Summary Chris Kerr  
5:30 Adjourn    
9/8/2011 (Thursday)    
8:30 Session 5 - Communication Issues on hybrid multi-core systems Mark Govett  
  Overlapping Computation and Communication for Advection on Hybrid Parallel Computers   Trey White
  Reworking Boundary Exchanges in HOMME for Many-Core Nodes   Ilene Carpenter
  Parallelization Strategy for Irregular Grids   Jim Rosinski
  Performance optimizations for running an NWP model on GPUs   Jacques Middlecoff
9:50 Discussion period 5    
10:20 Morning Break    
10:30 Session 6 - Compilers and Programming Models Will Sawyer  
  Accelerating weather models with PGI compilers   Dave Norton
  Multi-GPU management in HMPP 3.0   Francois Bodin
  Performance Evaluation of Programming Models for the Multicore Era   Nicholas Wright
  Porting S3D to the XK6 using OpenMP extensions for Accelerators   John Levesque
11:50 Discussion period 6    
12:20 Lunch    
1:30 Session 7 - Programming Approaches for Dynamical Cores Mark Govett  
  Rewrite of the COSMO Dynamical Core   Mueller / Gysi
  Experiences with the Finite-Volume Dynamical core and GEOS-5 on GPUs   Bill Putman
  Progress in Accelerating CAM-SE   Jeff Larkin
2:30 Afternoon Break    
3:00 Heterogeneous Programming for Hybrid CPU-GPU systems: Lessons learned from Comp Chemistry   Jeff Hammond
  Porting the ICON Non-hydrostatic Dynamical Solver to GPUs   Will Sawyer
3:40 Discussion period 7    
4:10 Workshop Summary and Future Plans Ilene Carpenter  
4:30 Adjourn