2015 MultiCore 5 Workshop- Agenda

 

Wednesday September 16, 2015

8:45 Workshop Introduction and Opening Remarks Rich Loft, NCAR

9:00

Session 1-Overviews and Strategies - Session Chair: Chris Kerr

9:00 - 9:20 Weather, Climate, and Earth System Modeling on Intel HPC Architectures: Practice, Experience, and Preparing for Future Generations –  (Slides) Richard Tran Mills, Intel
9:20 - 9:40 Update on GPU Acceleration of Earth System Models –  (Slides) Carl Ponder, NVIDIA
9:40 - 10:00 The CSIRO Accelerated Computing Strategy – (Slides) John Taylor, CSIRO
10:00 - 10:30 Morning Break  
10:30 - 10:50 Lessons Learned from Selected NERSC Application Readiness Application – (Slides) Helen He, LBL
10:50 - 11:10 Comparison of Parallelization Approaches for a Prototypical Climate Application Bryce Adelstein-Lelbach, LBNL
11:10 - 11:40 Session 1 Discussion  
11:40 - 1:00 Lunch  

1:00

Session 2: Tools and Techniques - Session Chair: Mike Ashworth

1:00 - 1:20 What's New in Intel Parallel Studio XE 2015 –  (Slides) Milind Girkar, Intel
1:20 - 1:40 To Thread or Not to Thread – (Slides) Jed Brown, CU
1:40 - 2:00 Adventures with Automatic Code Transformation – (Slides) Chris Maynard, UKMO
2:00 - 2:30 Afternoon Break  
2:30 - 2:50 PSyclone: A Code Generation and Optimization System for Finite Element and Finite Difference Codes – (Slides) Rupert Ford, STFC
2:50 - 3:10 Evaluating Performance and Scalability of Candidate Dynamical Cores for the Next Generation Global Prediction System at NOAA –  (Slides) John Michalakes, NOAA
3:10 - 3:30 KGEN in Practice –  (Slides) Youngsung Kim, NCAR
3:30 - 4:00 Session 2 Discussion  
4:00 Adjourn  

Thursday September 17, 2015

8:45

Session 3: US Climate Models - Session Chair: Ilene Carpenter

8:45 - 9:05 Experiences with CUDA and OpenACC from Porting ACME to GPUs –  (Slides) Matt Norman, ORNL
9:05 - 9:25 GPU Implementations of the CAM-SE Advection Limiter – (Slides) Irina Demeshko, SNL
9:25 - 9:50 Many-core Optimization of the Community Earth System Model – (Slides) John Dennis, NCAR
9:50 - 10:10 Morning Break  
10:10 - 10:30 Challenge in GFDL’s FMS Model Scaling: Communication – (Slides) Jeff Durachta, NOAA
10:30 -11:00 Session 3 discussion  
11:00 Group Photo  
11:15 – 12:30 Lunch – on your own in the NCAR Cafeteria  

12:30

Session 4: Weather Models - Session Chair: Mark Govett

12:30 - 12:50 Porting and Optimizing NCEP's GFS Physics Package for Unstructured Grids on Intel Xeon and Xeon Phi – (Slides) Jim Rosinski, NOAA
12:50 - 1:10 Optimzation of MPI Message Passing in a Multi-Core NWP Dynamical Core Running on NVIDIA GPUs Jacques Middlecoff, NOAA
1:10 – 1:30 Porting Operational Models to Multi- and Many-Core Architectures – (Slides) Ulrich Schaettler, DWD
1:30 - 2:00 Session 4 discussion  
2:00 - 2:30 Afternoon Break  

2:30

Session 5: Ocean Models - Session Chair: Chris Kerr

2:30 - 2:50 HBM Code Modernization - Current Status with a Focus on the Intel Xeon Phi Architecture – (Slides) Jacob Weismann Poulsen, DMI
2:50 - 3:10 Multi-core Efforts within the MPAS Framework with a Focus on the Ocean Core – (Slides) Doug Jacobsen, LANL
3:10 - 3:20 Performance Analysis of the PSyKAl Approach for a NEMO-based Benchmark – (Slides) Mike Ashworth, STFC
3:20 - 3:50 Session 5 Discussion  
3:50 - 4:00 Workshop Summary and discussion of plans for 2016  
4:00 Adjourn