|
|
|
|
Tuesday, September 13, 2016 |
|
|
8:00 – 8:45 |
Registration and Continental Breakfast |
|
8:45 – 9:00 |
Rich Loft, NCAR: Workshop Introduction |
|
Session 1: Updates and Strategies Session Chair: Chris Kerr |
|
|
9:00 – 9:30 |
Richard Tran Mills, Intel Corporation: Intel Update |
|
9:30 – 10:00 |
Milind Girkar, Intel Corporation : What's new Intel compilers |
|
10:00 – 10:30 |
Leslie Hart, NOAA/HPCC : NOAA's Software Engineering for Novel Architectures (SENA) Project |
|
10:30 – 11:00 |
Stan Posey, NVIDIA: NVIDIA update and Directions on GPU Acceleration for Earth System Models Carl Ponder, NVIDIA: Adaptation of the Weather Research and Forecasting Model to GPU's |
|
11:00 – 11:30 |
Group Photo and extended break |
|
11:30 – 12:30 |
Lunch in the ML Café, on your own, cash only |
|
Session 2: ESMs on Intel Xeon Phi Session Chair: Mark Govett |
|
|
12:30 – 1:00 |
Helen He, LBNL: NERSC Early KNL Experiences |
|
1:00 – 1:30 |
Hans Johansen, LBNL: Performance improvements and OpenMP design for a finite volume AMR dycore |
|
1:30 – 2:00 |
Azamat Mametjanov, ANL: Nested Threading and Vectorization with OpenMP in ACME |
|
2:00 – 2:30 |
John Dennis, NCAR: Increasing Thread Parallelism with the Higher Order Methods Modeling Environment |
|
2:30 – 3:00 |
Afternoon Break |
|
3:00 - 3:30 |
John Michalakes, Naval Research Laboratory: Experiences Programming and Optimizing for Knights Landing |
|
Session 3: ESMs on GPUs Session Chair: Hans Johansen |
|
|
3:30 - 4:00 |
Raymond Menzel, NOAA/GFDL: Challenges of offloading radiative heating calculations using OpenACC and OpenMP accelerator device directives in the CM2.5 FLOR Earth System Model |
|
4:00 – 4:30 |
Mark Govett, NOAA/ESRL: GPU Parallelization of the FV3 Model |
|
4:30 – 5:00 |
Ulrich Schättler, Deutscher Wetterdienst: Accelerating Work at DWD |
|
|
|
|
Wednesday, September 14, 2016 |
|
|
8:30 – 9:00 |
Arrival, Continental Breakfast |
|
Session 4: Towards Exascale Session Chair: Ilene Carpenter |
|
|
9:00 – 9:30 |
Daniel S. Abdi, Naval Postgraduate School: Towards Exascale Computing with the Atmospheric Model NUMA |
|
9:30 – 10:00 |
Mike Ashworth, Science and Technology Facilities Council: PSyclone: a code generation and optimisation system for finite element and finite difference codes |
|
10:00 – 10:30 |
Chris Maynard, Met Office: Many more cores: Weather and Climate models beyond 100k cores |
|
10:30 – 11:00 |
Morning Break |
|
Session 5: ESMs on Intel Xeon Phi Session Chair: Mike Ashworth |
|
|
11:00 – 11:30 |
Jed Brown, CU Boulder: Design Considerations for Latency and Throughput on KNL |
|
11:30-12:00 |
Jim Rosinski, NOAA/ESRL: Preparing atmospheric modeling codes for the latest generation MIC architecture (KNL) |
|
12:00 – 1:00 |
Lunch in ML Café, on your own, cash only |
|
1:00 – 1:30 |
Lynd Stringer, NOAA : Analysis of FV3's subroutine xppm0 |
|
1:30 – 2:00 |
Marcus Wagner, Cray Inc: Performance Scaling Studies of HOMME Over Multiple KNL Nodes |
|
2:00 – 2:30 |
Youngsung Kim, NCAR: Application of KGen and KGen-generated kernel |
|
2:30 – 3:00 |
Afternoon Break |
|
Session 6: ESMs on GPUs Session Chair: Chris Kerr |
|
|
3:00 – 3:30 |
Venkatramani Balaji, Princeton University: On the Application of GPU Hardware for Line by Line Radiation Computations |
|
3:30 – 4:00 |
Ji-Sun Kang, KISTI: Development of heterogeneous codes of MPAS physics using OpenACC directives: Preliminary results from WSM6 microphysics scheme |
|
Session 7: Wrap-up Session Chair: Chris Kerr |
|
|
4:00 – 4:15 |
Chris Kerr, NCAR Consultant: Wrap-up and workshop concludes |
*Please note that not all presentations can be posted due to legal concerns. Thank you for understanding.
