This title appears in the Scientific Report :
2023
Please use the identifier:
http://dx.doi.org/10.34734/FZJ-2023-04519 in citations.
Scalable Lattice Boltzmann Leaps to Exascale
Scalable Lattice Boltzmann Leaps to Exascale
Lattice Boltzmann methods (LBM) are trustworthy alternatives to conventionalCFD, showing roughly an order of magnitude performance advantage thanNavier-Stokes approaches in comparable scenarios. The S CALABLE projectbrings together industrial and academic partners to create a new frameworkfor an ind...
Saved in:
Personal Name(s): | Badwaik, Jayesh (Corresponding author) |
---|---|
Contributing Institute: |
Jülich Supercomputing Center; JSC |
Imprint: |
2023
|
DOI: |
10.34734/FZJ-2023-04519 |
Conference: | ISC High Performance 2024, Hamburg (Germany), 2023-05-23 - 2023-05-26 |
Document Type: |
Poster |
Research Program: |
SCAlable LAttice Boltzmann Leaps to Exascale Cross-Domain Algorithms, Tools, Methods Labs (ATMLs) and Research Groups |
Link: |
OpenAccess |
Publikationsportal JuSER |
Lattice Boltzmann methods (LBM) are trustworthy alternatives to conventionalCFD, showing roughly an order of magnitude performance advantage thanNavier-Stokes approaches in comparable scenarios. The S CALABLE projectbrings together industrial and academic partners to create a new frameworkfor an industrial LBM-based computational fluid dynamics (CFD) solver whichcan achieve high performance, scalability, and energy efficiency.In the context of EuroHPC, LBM is especially well suited to exploit advancedsupercomputer architectures through vectorization, accelerators, and massiveparallelization. To achieve its goals, SCALABLE aims to transfer leading edgeperformance technology between waLBerla and LaBS, thus breaking the silosbetween the worlds of scientific computing and physical flow modelling. |