Extreme Scale Plasma Turbulence Simulations on Top Supercomputers Worldwide

Journal Article · Tue Nov 01 00:00:00 EDT 2016 · International Conference for High Performance Computing, Networking, Storage and Analysis (Online)

DOI:https://doi.org/10.1109/SC.2016.42· OSTI ID:1379775

Tang, William ^[1]; Wang, Bei ^[2]; Ethier, Stephane ^[3]; Kwasniewski, Grzegorz ^[4]; Hoefler, Torsten ^[4]; Ibrahim, Khaled Z. ^[5]; Madduri, Kamesh ^[6]; Williams, Samuel ^[5]; Oliker, Leonid ^[5]; Rosales-Fernandez, Carlos ^[7]; Williams, Tim ^[8]

---

1. Princeton Univ., NJ (United States). Princeton Inst. for Computational Science and Engineering; Princeton Plasma Physics Lab. (PPPL), Princeton, NJ (United States)
2. Princeton Univ., NJ (United States). Princeton Inst. for Computational Science and Engineering
3. Princeton Plasma Physics Lab. (PPPL), Princeton, NJ (United States)
4. ETH Zurich (Switzerland). Computer Science Dept.
5. Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States). Computational Research Division
6. Pennsylvania State Univ., University Park, PA (United States). Computer Science and Engineering Dept.
7. Univ. of Texas, Austin, TX (United States). Texas Advanced Computing Center
8. Argonne National Lab. (ANL), Argonne, IL (United States). Argonne Leadership Comuting Facility

The goal of the extreme scale plasma turbulence studies described in this paper is to expedite the delivery of reliable predictions on confinement physics in large magnetic fusion systems by using world-class supercomputers to carry out simulations with unprecedented resolution and temporal duration. This has involved architecture-dependent optimizations of performance scaling and addressing code portability and energy issues, with the metrics for multi-platform comparisons being 'time-to-solution' and 'energy-to-solution'. Realistic results addressing how confinement losses caused by plasma turbulence scale from present-day devices to the much larger $25 billion international ITER fusion facility have been enabled by innovative advances in the GTC-P code including (i) implementation of one-sided communication from MPI 3.0 standard; (ii) creative optimization techniques on Xeon Phi processors; and (iii) development of a novel performance model for the key kernels of the PIC code. Our results show that modeling data movement is sufficient to predict performance on modern supercomputer platforms.

• Research Organization: Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)
• Sponsoring Organization: USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
• Grant/Contract Number: AC02-05CH11231; AC02-06CH11357; AC02-09CH11466; AC02-00OR22725
• OSTI ID: 1379775
• Journal Information: International Conference for High Performance Computing, Networking, Storage and Analysis (Online), Vol. 2017; Conference: International Conference for High Performance Computing, Networking, Storage and Analysis, Salt Lake City, UT (United States), 13-18 Nov 2016; ISSN 2167-4337
• Publisher: IEEE
• Country of Publication: United States
• Language: English

Related Subjects

97 MATHEMATICS AND COMPUTING