TY - GEN
T1 - Towards extreme-scale simulations with next-generation trilinos
T2 - 28th IEEE International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014
AU - Lin, Paul
AU - Bettencourt, Matthew
AU - Domino, Stefan
AU - Fisher, Travis
AU - Hoemmen, Mark
AU - Hu, Jonathan
AU - Phipps, Eric
AU - Prokopenko, Andrey
AU - Rajamanickam, Sivasankaran
AU - Siefert, Christopher
AU - Cyr, Eric
AU - Kennon, Stephen
N1 - Publisher Copyright:
© 2014 IEEE.
PY - 2014/11/27
Y1 - 2014/11/27
N2 - Trilinos is an object-oriented software framework for the solution of large-scale, complex multi-physics engineering and scientific problems. While the original version of Trilinos was designed for highly scalable solutions for large problems, the need for increasingly higher fidelity simulations has pushed the problem sizes beyond what could have been envisioned two decades ago. When problem sizes exceed a billion elements even highly scalable applications and solver stacks require a complete revision. The next-generation Trilinos employs C++ templates in order to solve arbitrarily large problems and enable extreme-scale simulations. We present a case study that involves integration of Trilinos with an engineering application (Sierra low Mach module/Nalu), involving the simulation of low Mach fluid flow for problems of size up to nine billion elements. Through the use of improved algorithms and better software engineering practices, we demonstrate good weak scaling for the matrix assembly and solve for the engineering application for up to a nine billion element fluid flow large eddy simulation (LES) problem on unstructured meshes with a 27 billion row matrix on 131,072 cores of a Cray XE6 platform.
AB - Trilinos is an object-oriented software framework for the solution of large-scale, complex multi-physics engineering and scientific problems. While the original version of Trilinos was designed for highly scalable solutions for large problems, the need for increasingly higher fidelity simulations has pushed the problem sizes beyond what could have been envisioned two decades ago. When problem sizes exceed a billion elements even highly scalable applications and solver stacks require a complete revision. The next-generation Trilinos employs C++ templates in order to solve arbitrarily large problems and enable extreme-scale simulations. We present a case study that involves integration of Trilinos with an engineering application (Sierra low Mach module/Nalu), involving the simulation of low Mach fluid flow for problems of size up to nine billion elements. Through the use of improved algorithms and better software engineering practices, we demonstrate good weak scaling for the matrix assembly and solve for the engineering application for up to a nine billion element fluid flow large eddy simulation (LES) problem on unstructured meshes with a 27 billion row matrix on 131,072 cores of a Cray XE6 platform.
KW - Extreme-scale simulations
KW - Solver library
KW - Trilinos
KW - Vertical integration
UR - http://www.scopus.com/inward/record.url?scp=84918771311&partnerID=8YFLogxK
U2 - 10.1109/IPDPSW.2014.166
DO - 10.1109/IPDPSW.2014.166
M3 - Conference contribution
AN - SCOPUS:84918771311
T3 - Proceedings - IEEE 28th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014
SP - 1485
EP - 1494
BT - Proceedings - IEEE 28th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2014
PB - IEEE Computer Society
Y2 - 19 May 2014 through 23 May 2014
ER -