210 likes | 312 Views
Integrating Trilinos Solvers to SEAM code. Dagoberto A.R. Justo – UNM Tim Warburton – UNM Bill Spotz – Sandia. SEAM (NCAR). Trilinos (Sandia Lab). AztecOO Epetra Nox Ifpack PETSc Komplex. Spectral Element Atmospheric Method. AztecOO. Solvers CG, CGS, BICGStab, GMRES, Tfqmr
E N D
Integrating Trilinos Solvers to SEAM code Dagoberto A.R. Justo – UNM Tim Warburton – UNM Bill Spotz – Sandia
SEAM (NCAR) Trilinos (Sandia Lab) • AztecOO • Epetra • Nox • Ifpack • PETSc • Komplex Spectral Element Atmospheric Method
AztecOO • Solvers • CG, CGS, BICGStab, GMRES, Tfqmr • Preconditioners • Diagonal Jacobi, Least Square, Neumann, Domain Decomposition, Symmetric Gauss-Seidel • Matrix Free implementation • C++ (Fortran interface) • MPI
Implementation A Z T E C SEAM CODE . . . Pcg_solver . . (F90) Pcg_solver . . Aztec_solvers( ) . (F90) Sub Aztec_solvers . AZ_Iterate( ) (C) Matrix_vector_C (C) Prec_Jacobi_C (C) Matrix_vector . (F90) Prec_Jacobi . (F90)
Machines used • Pentium III Notebook (serial) • Linux, LAM-MPI, Intel Compilers • Los Lobos at HPC@UNM • Linux Cluster • 256 nodes • IBM Pentium III 750 MHz, 256 KB L2 Cache, 1 Gb RAM • Portland Group compiler • MPICH for Myrinet interconnections
Graphical Results from SEAM Mass Energy
Speed Up • From 1 to 160 processors. • Time of Simulation 144 time iterations x 300 s = 12 h simulation • Verify results using mass, energy,… • (Different result for 1 proc)
Speed Up – SEAM+Aztecbest: cgs solver + Least Square preconditioner
Conclusions &Suggested Future Efforts • SEAM+Aztec works! • SEAM+Aztec is 2x slower difference in CG algorithms SEAM+Aztec time-iteration is 50% slower 0.1% of time lost in calls, preparation for Aztec. • More time better tune-up. • Domain decomposition Preconditioners
Conclusions &Suggested Future Efforts • SEAM + Aztec works! • More time better tune-up.