The parallel inhouse research CFD code SCH3D



Universitat Politècnica de Catalunya (UPC)

 Highorder algorithms for structured meshes
 Large scale DNS and LES of incompressible flows
 High efficiency on thousands of CPUs of a supercomputer
 High performance scalable Poisson solvers
 Flexible configuration to fit both small "Beowulf" clusters and supercomputers

Authors: X. F. Trias, A. Gorobets, M. Soria

Mathematical basis
 Highorder robust symmetry preserving discretization
 Explicit time integration
 Fractional step projection method for pressurevelocity coupling
 Innovative LES models
 Scalable Poisson solvers


Implementation
 The code is written on C language
 MPI parallelization
 Objectoriented structure

Development directions
 LES models (Trias)
 Poisson solvers (Gorobets, Trias)
 Improvements of parallel algorithms, hybrid MPI+OpenMP approach (Gorobets)

Current applications
 Largescale DNS of model problems. More...
 Development and verification of LES models

Publications:
 A. Gorobets, F. X. Trias, M. Soria and A. Oliva, “A scalable parallel Poisson solver for threedimensional
problems with one periodic direction”, Computers & Fluids journal, 39 (2010) 525538, Elsevier
 A.V.Gorobets, “Scalable algorithm for incompressible flow simulation on parallel computer systems”, Mathematical modeling, vol. 19, number 10, pp. 105128, 2007
 F. X. Trias, A. Gorobets, M. Soria and A. Oliva, “A scalable ACMbased Poisson solver for DNS of Turbulent Flows on supercomputers”, Parallel CFD 2008, Lyon (France), May 2008
 R. W. C. P. Verstappen and A. E. P. Veldman. "SymmetryPreserving Discretiza
tion of Turbulent Flow", Journal of Computational Physics, 187:343{368, May
2003.

