Coupling Exascale Multiphysics Applications: Methods and Lessons Learned

Conference · 01 October 2018

DOI:https://doi.org/10.1109/eScience.2018.00133· OSTI ID:1545215

Choi, Jong Youl ^[1]; Chang, C.S. ^[2]; Dominski, Julien ^[2]; Klasky, Scott A. ^[1]; Merlo, Gabriele ^[3]; Suchyta, Eric D. ^[1]; Ainsworth, Mark ^[1]; Allen, Bryce ^[4]; Cappello, Franck ^[4]; Churchill, Michael ^[2]; Davis, Philip ^[5]; Di, Sheng ^[4]; Eisenhauer, Greg ^[6]; Ethier, Stephane ^[2]; Foster, Ian ^[4]; Geveci, Berk ^[7]; Guo, Hanqi ^[4]; Huck, Kevin ^[8]; Jenko, Frank ^[9]; Kim, Mark B. ^[1] ; Kress, James M. ^[1]; Ku, Seung-Hoe ^[2]; Liu, Qing Gary ^[1]; Logan, Jeremy S. ^[1]; Malony, Allen ^[8]; Mehta, Kshitij V. ^[1]; Moreland, Kenneth ^[10]; Munson, Todd ^[4]; Parashar, Manish ^[11]; Peterka, Tom ^[4]; Podhorszki, Norbert ^[1]; Pugmire, Dave ^[1]; Tugluk, Ozan ^[12]; Wang, Ruonan ^[1]; Whitney, Ben ^[12]; Wolf, Matthew D. ^[1]; Wood, Chad ^[8]

---

1. ORNL
2. Princeton Plasma Physics Laboratory (PPPL)
3. University of Texas, Austin
4. Argonne National Laboratory (ANL)
5. Rutgers University, Newark, NJ
6. Georgia Institute of Technology, Atlanta
7. Kitware
8. University of Oregon
9. University of California, Los Angeles
10. Sandia National Laboratories (SNL)
11. Rutgers University
12. Brown University

With the growing computational complexity of science and the complexity of new and emerging hardware, it is time to re-evaluate the traditional monolithic design of computational codes. One new paradigm is constructing larger scientific computational experiments from the coupling of multiple individual scientific applications, each targeting their own physics, characteristic lengths, and/or scales. We present a framework constructed by leveraging capabilities such as in-memory communications, workflow scheduling on HPC resources, and continuous performance monitoring. This code coupling capability is demonstrated by a fusion science scenario, where differences between the plasma at the edges and at the core of a device have different physical descriptions. This infrastructure not only enables the coupling of the physics components, but it also connects in situ or online analysis, compression, and visualization that accelerate the time between a run and the analysis of the science content. Results from runs on Titan and Cori are presented as a demonstration.

• Research Organization: Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States). Oak Ridge Leadership Computing Facility (OLCF); Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States)
• Sponsoring Organization: USDOE Office of Science (SC)
• DOE Contract Number: AC05-00OR22725
• OSTI ID: 1545215
• Country of Publication: United States
• Language: English