TY - JOUR
T1 - A Collision-based Domain Decomposition scheme for large-scale depletion with the Serpent 2 Monte Carlo code
AU - García, Manuel
AU - Leppänen, Jaakko
AU - Sanchez-Espinoza, Victor
N1 - Funding Information:
This work was performed on the computational resource ForHLR II funded by the Ministry of Science, Research and the Arts Baden-Württemberg and DFG (”Deutsche Forschungsgemeinschaft”).
Funding Information:
This work was done within the McSAFE project which is receiving funding from the Euratom research and training programme 2014–2018 under grant agreement No 755097.
Funding Information:
This work was done within the McSAFE project which is receiving funding from the Euratom research and training programme 2014?2018 under grant agreement No 755097. This work was performed on the computational resource ForHLR II funded by the Ministry of Science, Research and the Arts Baden-W?rttemberg and DFG (?Deutsche Forschungsgemeinschaft?).
Publisher Copyright:
© 2020 Elsevier Ltd
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2021/3
Y1 - 2021/3
N2 - In the framework of the EU Horizon 2020 McSAFE project, high-fidelity multiphysics capabilities are being developed to carry out large-scale burnup calculations for Light Water Reactors. As part of this effort, the Serpent 2 Monte Carlo code has been coupled to thermalhydraulics and fuel-performance codes, with the final objective of performing fully coupled full-core pin-by-pin simulations. To enable memory scalability, needed for these massive problems, a Collision-based Domain Decomposition (CDD) scheme has been implemented in Serpent 2. The methodology is based on data decomposition for burnable materials and a domain decomposition particle-tracking algorithm, and is shown here to provide the required memory scalability and computational performance, with up to 50% speedup efficiency at 5,120 cores. The application of the CDD feature is demonstrated in a pin-by-pin depletion calculation for a Pre-Konvoi PWR reactor.
AB - In the framework of the EU Horizon 2020 McSAFE project, high-fidelity multiphysics capabilities are being developed to carry out large-scale burnup calculations for Light Water Reactors. As part of this effort, the Serpent 2 Monte Carlo code has been coupled to thermalhydraulics and fuel-performance codes, with the final objective of performing fully coupled full-core pin-by-pin simulations. To enable memory scalability, needed for these massive problems, a Collision-based Domain Decomposition (CDD) scheme has been implemented in Serpent 2. The methodology is based on data decomposition for burnable materials and a domain decomposition particle-tracking algorithm, and is shown here to provide the required memory scalability and computational performance, with up to 50% speedup efficiency at 5,120 cores. The application of the CDD feature is demonstrated in a pin-by-pin depletion calculation for a Pre-Konvoi PWR reactor.
KW - Collision-based Domain Decomposition
KW - High Performance Computing
KW - Pin-level depletion
KW - Serpent 2
UR - http://www.scopus.com/inward/record.url?scp=85097134976&partnerID=8YFLogxK
U2 - 10.1016/j.anucene.2020.108026
DO - 10.1016/j.anucene.2020.108026
M3 - Article
AN - SCOPUS:85097134976
VL - 152
JO - Annals of Nuclear Energy
JF - Annals of Nuclear Energy
SN - 0306-4549
M1 - 108026
ER -