Parallel Computational Fluid Dynamics
Documento
Informações da Disciplina
Código
COC898
Título
Parallel Computational Fluid Dynamics
Ementa
• Introduction to Parallel CFD
• Supercomputer anatomy: hardware and software
• Metrics: speedup, scalability (weak and strong), load balancing
• Profiling tools (VTune, TAU)
• Challenges: memory, power, I/O
• Anatomy of a CFD tool
• The discrete problem (FEM, FDM, FVM)
• Mesh generation, adaptivity and refinement
• Partitioning and load balancing (static and dynamic)
• Data structures: assembly, CSR, EBE, EDE
• Distributed and shared memory parallelism
• Time loop, non-linear and linear problems
• I/O and Visualization
• Workflows: a posteriori, in-situ, local and remote
• Methods: scalar and vector fields, filters, pipelines
Bibliografia
• Houzeaux, G. [et al.]. High-Performance Computing: Do’s and Dont’s. A: "Computational Fluid Dynamics - Basic Instruments and Applications in Science". InTech, 2018, p. 3-41.
• Dongarra, J. [et al.], “Sourcebook of Parallel Computing”, The Morgan Kaufmann Series in Computer Architecture and Design, 1st Edition, 2002
• Tezduyar, T. “Finite Element in Fluids Series”, Lecture Notes of the Shortcourse on Finite Elements in Fluids”, Computational Mechanics Division - Vol. 99-77, Japan Society of Mechanical Engineers, Japan, 1999
Créditos
3.0/45h
Professor
Alvaro Luis Gayoso de Azeredo Coutinho
Renato Nascimento Elias
Idioma
Inglês