Parallel sparse linear solvers and applications in CFD
|
|
- Tracy Day
- 5 years ago
- Views:
Transcription
1 Parallel sparse linear solvers and applications in CFD Jocelyne Erhel Joint work with Désiré Nuentsa Wakam () and Baptiste Poirriez () SAGE team, Inria Rennes, France journée Calcul Intensif Distribué dans l Industrie, Université Paris 13, 22 janvier / 21
2 Outline Solver interface Interface to direct and iterative solvers: MUMPS, SuperLU Dist, Hypre, Petsc, parms,etc SLSI [Nuentsa Wakam et al 2010] available on demand System solver in H2OLab platform [Erhel et al 2009] Application to CFD problems (m): a Krylov method for general matrices combining Domain Decomposition and deflation : a Krylov method for SPD matrices combining Domain Decomposition and deflation 2 / 21
3 Preconditioned Ax = b, A R n n x, b R n B = AM 1 (m): a Krylov subspace method [Saad and Schultz 1986, Meurant s book 1999, Saad s book 2003, Simoncini and Szyld 2007, Erhel 2011,...] Fix x 0, then r 0 = b Ax 0 K m(b, r 0) = span{r 0, Br 0,... B m 1 r 0} D A Find x m x 0 + K m(b, r 0) such that r m 2 = b Bx m 2 = min u x0 +Km(B,r 0 ) b Bu 2 Building blocks of Initial step: choose x 0, compute r 0 First step: generate an orthonormal basis V m+1 = [v 0,..., v m] of K m+1(b, r 0) such that v 0 = r 0/β, β = r 0 2, BV m = V m+1 H m Second step: approximate solution x m = x 0 + M 1 V my m r m = r 0 BV my m = V m+1(βe 1 H my m) y m = min y R m βe 1 H my 2 3 / 21
4 ... practical Arnoldi process 1: v 0 = r 0 / r 0 2 2: for k = 0,... do 3: p = Bv k 4: for i = 1 : k do 5: h ik = vi T p 6: p = p h ik v i 7: end for 8: h k+1,k = p 2 9: v k+1 = p/h k+1,k 10: end for BV m = V m+1 H m Granularity in parallel algorithms Communication-avoiding strategies Generate the basis vectors [Reichel 1990, Bai et al 1994] Orthogonalize the basis [De Sturler 1994, Erhel 1995, Sidje 1997] Improve the strategy [Hoemmen 2010, Demmel et al 2011] Preconditioning use multilevel methods to deal with large systems Schwarz preconditioning [Atenekeng Kahou et al 2007, Dufaud+Tromeur-Dervout 2010, Giraud+Haidar 2009, Smith et al s book 1996,...] Filtering and Schur complement [Li et al 2003, Grigori et al 2011] Multilevel parallelism [Nuentsa Wakam et al 2011, Giraud et al 2010,...] Complexity and stagnation with restarted (m) Use deflation to recover possible loss of information Deflation by preconditioning [Erhel et al 1996, Burrage et al 1998, Baglama et al 1998,...] Deflation by augmented basis [Morgan 1995, Morgan 2002,...] D A Work in the team SAGE Combine communication-avoiding... and Deflation... and domain decomposition preconditioners [Nuentsa Wakam 2011, Nuentsa Wakam+Erhel+Gropp 2013, Nuentsa Wakam+Pacull 2013, Nuentsa Wakam+Erhel 2014] 4 / 21
5 ... practical Arnoldi process 1: v 0 = r 0 / r 0 2 2: for k = 0,... do 3: p = Bv k 4: for i = 1 : k do 5: h ik = vi T p 6: p = p h ik v i 7: end for 8: h k+1,k = p 2 9: v k+1 = p/h k+1,k 10: end for BV m = V m+1 H m Granularity in parallel algorithms Communication-avoiding strategies Generate the basis vectors [Reichel 1990, Bai et al 1994] Orthogonalize the basis [De Sturler 1994, Erhel 1995, Sidje 1997] Improve the strategy [Hoemmen 2010, Demmel et al 2011] Preconditioning use multilevel methods to deal with large systems Schwarz preconditioning [Atenekeng Kahou et al 2007, Dufaud+Tromeur-Dervout 2010, Giraud+Haidar 2009, Smith et al s book 1996,...] Filtering and Schur complement [Li et al 2003, Grigori et al 2011] Multilevel parallelism [Nuentsa Wakam et al 2011, Giraud et al 2010,...] Complexity and stagnation with restarted (m) Use deflation to recover possible loss of information Deflation by preconditioning [Erhel et al 1996, Burrage et al 1998, Baglama et al 1998,...] Deflation by augmented basis [Morgan 1995, Morgan 2002,...] D A Work in the team SAGE Combine communication-avoiding... and Deflation... and domain decomposition preconditioners [Nuentsa Wakam 2011, Nuentsa Wakam+Erhel+Gropp 2013, Nuentsa Wakam+Pacull 2013, Nuentsa Wakam+Erhel 2014] 4 / 21
6 ... practical Arnoldi process 1: v 0 = r 0 / r 0 2 2: for k = 0,... do 3: p = Bv k 4: for i = 1 : k do 5: h ik = vi T p 6: p = p h ik v i 7: end for 8: h k+1,k = p 2 9: v k+1 = p/h k+1,k 10: end for BV m = V m+1 H m Granularity in parallel algorithms Communication-avoiding strategies Generate the basis vectors [Reichel 1990, Bai et al 1994] Orthogonalize the basis [De Sturler 1994, Erhel 1995, Sidje 1997] Improve the strategy [Hoemmen 2010, Demmel et al 2011] Preconditioning use multilevel methods to deal with large systems Schwarz preconditioning [Atenekeng Kahou et al 2007, Dufaud+Tromeur-Dervout 2010, Giraud+Haidar 2009, Smith et al s book 1996,...] Filtering and Schur complement [Li et al 2003, Grigori et al 2011] Multilevel parallelism [Nuentsa Wakam et al 2011, Giraud et al 2010,...] Complexity and stagnation with restarted (m) Use deflation to recover possible loss of information Deflation by preconditioning [Erhel et al 1996, Burrage et al 1998, Baglama et al 1998,...] Deflation by augmented basis [Morgan 1995, Morgan 2002,...] D A Work in the team SAGE Combine communication-avoiding... and Deflation... and domain decomposition preconditioners [Nuentsa Wakam 2011, Nuentsa Wakam+Erhel+Gropp 2013, Nuentsa Wakam+Pacull 2013, Nuentsa Wakam+Erhel 2014] 4 / 21
7 ... practical Arnoldi process 1: v 0 = r 0 / r 0 2 2: for k = 0,... do 3: p = Bv k 4: for i = 1 : k do 5: h ik = vi T p 6: p = p h ik v i 7: end for 8: h k+1,k = p 2 9: v k+1 = p/h k+1,k 10: end for BV m = V m+1 H m Granularity in parallel algorithms Communication-avoiding strategies Generate the basis vectors [Reichel 1990, Bai et al 1994] Orthogonalize the basis [De Sturler 1994, Erhel 1995, Sidje 1997] Improve the strategy [Hoemmen 2010, Demmel et al 2011] Preconditioning use multilevel methods to deal with large systems Schwarz preconditioning [Atenekeng Kahou et al 2007, Dufaud+Tromeur-Dervout 2010, Giraud+Haidar 2009, Smith et al s book 1996,...] Filtering and Schur complement [Li et al 2003, Grigori et al 2011] Multilevel parallelism [Nuentsa Wakam et al 2011, Giraud et al 2010,...] Complexity and stagnation with restarted (m) Use deflation to recover possible loss of information Deflation by preconditioning [Erhel et al 1996, Burrage et al 1998, Baglama et al 1998,...] Deflation by augmented basis [Morgan 1995, Morgan 2002,...] D A Work in the team SAGE Combine communication-avoiding... and Deflation... and domain decomposition preconditioners [Nuentsa Wakam 2011, Nuentsa Wakam+Erhel+Gropp 2013, Nuentsa Wakam+Pacull 2013, Nuentsa Wakam+Erhel 2014] 4 / 21
8 ... practical Arnoldi process 1: v 0 = r 0 / r 0 2 2: for k = 0,... do 3: p = Bv k 4: for i = 1 : k do 5: h ik = vi T p 6: p = p h ik v i 7: end for 8: h k+1,k = p 2 9: v k+1 = p/h k+1,k 10: end for BV m = V m+1 H m Granularity in parallel algorithms Communication-avoiding strategies Generate the basis vectors [Reichel 1990, Bai et al 1994] Orthogonalize the basis [De Sturler 1994, Erhel 1995, Sidje 1997] Improve the strategy [Hoemmen 2010, Demmel et al 2011] Preconditioning use multilevel methods to deal with large systems Schwarz preconditioning [Atenekeng Kahou et al 2007, Dufaud+Tromeur-Dervout 2010, Giraud+Haidar 2009, Smith et al s book 1996,...] Filtering and Schur complement [Li et al 2003, Grigori et al 2011] Multilevel parallelism [Nuentsa Wakam et al 2011, Giraud et al 2010,...] Complexity and stagnation with restarted (m) Use deflation to recover possible loss of information Deflation by preconditioning [Erhel et al 1996, Burrage et al 1998, Baglama et al 1998,...] Deflation by augmented basis [Morgan 1995, Morgan 2002,...] D A Work in the team SAGE Combine communication-avoiding... and Deflation... and domain decomposition preconditioners [Nuentsa Wakam 2011, Nuentsa Wakam+Erhel+Gropp 2013, Nuentsa Wakam+Pacull 2013, Nuentsa Wakam+Erhel 2014] 4 / 21
9 Communication-avoiding strategy: Newton basis building blocks Initial step: run one cycle of (m) and compute shifts for the Newton basis First step: build a basis K m+1 = [k 0, k 1,..., k m] of the Krylov subspace K m+1 (B, r 0 ) such that BK m = K m+1 T m D A Second step: compute an orthonormal basis of K m+1 (B, r 0 ) Compute the QR factorization K m+1 = V m+1 R m+1 RODDEC [Sidje 1997, Erhel 1995] or TSQR [Demmel et al 2011] BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m }{{} Hm Third step: approximate solution x m = x 0 + M 1 V my m r m = r 0 BK my m = V m+1 (βe 1 H my m) with β = r 0 2 y m = min y R m βe 1 H my 2 5 / 21
10 Communication-avoiding strategy: Newton basis building blocks Initial step: run one cycle of (m) and compute shifts for the Newton basis First step: build a basis K m+1 = [k 0, k 1,..., k m] of the Krylov subspace K m+1 (B, r 0 ) such that BK m = K m+1 T m D A Second step: compute an orthonormal basis of K m+1 (B, r 0 ) Compute the QR factorization K m+1 = V m+1 R m+1 RODDEC [Sidje 1997, Erhel 1995] or TSQR [Demmel et al 2011] BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m }{{} Hm Third step: approximate solution x m = x 0 + M 1 V my m r m = r 0 BK my m = V m+1 (βe 1 H my m) with β = r 0 2 y m = min y R m βe 1 H my 2 5 / 21
11 Communication-avoiding strategy: Newton basis building blocks Initial step: run one cycle of (m) and compute shifts for the Newton basis First step: build a basis K m+1 = [k 0, k 1,..., k m] of the Krylov subspace K m+1 (B, r 0 ) such that BK m = K m+1 T m D A Second step: compute an orthonormal basis of K m+1 (B, r 0 ) Compute the QR factorization K m+1 = V m+1 R m+1 RODDEC [Sidje 1997, Erhel 1995] or TSQR [Demmel et al 2011] BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m }{{} Hm Third step: approximate solution x m = x 0 + M 1 V my m r m = r 0 BK my m = V m+1 (βe 1 H my m) with β = r 0 2 y m = min y R m βe 1 H my 2 5 / 21
12 Communication-avoiding strategy: Newton basis building blocks Initial step: run one cycle of (m) and compute shifts for the Newton basis First step: build a basis K m+1 = [k 0, k 1,..., k m] of the Krylov subspace K m+1 (B, r 0 ) such that BK m = K m+1 T m D A Second step: compute an orthonormal basis of K m+1 (B, r 0 ) Compute the QR factorization K m+1 = V m+1 R m+1 RODDEC [Sidje 1997, Erhel 1995] or TSQR [Demmel et al 2011] BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m }{{} Hm Third step: approximate solution x m = x 0 + M 1 V my m r m = r 0 BK my m = V m+1 (βe 1 H my m) with β = r 0 2 y m = min y R m βe 1 H my 2 5 / 21
13 combined with Domain Decomposition Main steps with Domain Decomposition preconditioning Partition the weighted graph of the matrix in parallel with PARMETIS. Redistribute the matrix and right-hand-side according to the PARMETIS partitioning. Perform a parallel iterative row and column scaling on the matrix and the right-hand side vector [Amestoy et al, 2008]. Define the overlap between the submatrices for the additive Schwarz preconditioner. D A M 1 RAS = Setup the submatrices (ILU or LU factorization). Solve iteratively the preconditioned system using. D (R 0 k )T (A δ k ) 1 R δ k k=1 6 / 21
14 Deflation strategies Restarted (m) x m = x 0 + M 1 V my m where y m minimizes r m 2 The convergence rate depends on the spectral distribution in B Smallest eigenvalues slow down the convergence Deflation occurs when the Krylov subspace is large enough With restarting : loss of spectral information, risk of stalling D A Accelerating the restarted [Simoncini and Szyld, 2007] Approximate the smallest eigenvalues and the associated invariant subspace U r Explicit deflation technique [Erhel et al 1996; Burrage et al 1998; Moriya et al 2000 ]: B M 1 x = b with M 1 = (I n + U r ( λ n T 1 I r )Ur T and T = Ur T BUr Augmented techniques [Morgan 2000, 2002, Giraud et al, 2010]: x m x 0 + span{u r } + K m(b, r 0 ) 7 / 21
15 Deflation strategies Restarted (m) x m = x 0 + M 1 V my m where y m minimizes r m 2 The convergence rate depends on the spectral distribution in B Smallest eigenvalues slow down the convergence Deflation occurs when the Krylov subspace is large enough With restarting : loss of spectral information, risk of stalling D A Accelerating the restarted [Simoncini and Szyld, 2007] Approximate the smallest eigenvalues and the associated invariant subspace U r Explicit deflation technique [Erhel et al 1996; Burrage et al 1998; Moriya et al 2000 ]: B M 1 x = b with M 1 = (I n + U r ( λ n T 1 I r )Ur T and T = Ur T BUr Augmented techniques [Morgan 2000, 2002, Giraud et al, 2010]: x m x 0 + span{u r } + K m(b, r 0 ) 7 / 21
16 D: with adaptive preconditioning deflation D(m, r) Perform one cycle of restarted (m) and compute shifts for the Newton basis compute U r, a basis of a coarse subspace Build M 1 r I n + U r ( λ n T 1 I r )U T r, T = UT r BUr Apply (m) to B M 1 r At each restart, update r and the basis U r D A Adaptive D(m,r) Detect a potential slow convergence [Sosonkina et al 1998] Switch to D(m,r) only if necessary [Nuentsa Wakam et al 2013] 8 / 21
17 D: with adaptive preconditioning deflation D(m, r) Perform one cycle of restarted (m) and compute shifts for the Newton basis compute U r, a basis of a coarse subspace Build M 1 r I n + U r ( λ n T 1 I r )U T r, T = UT r BUr Apply (m) to B M 1 r At each restart, update r and the basis U r D A Adaptive D(m,r) Detect a potential slow convergence [Sosonkina et al 1998] Switch to D(m,r) only if necessary [Nuentsa Wakam et al 2013] 8 / 21
18 D: with adaptive preconditioning deflation D(m, r) Perform one cycle of restarted (m) and compute shifts for the Newton basis compute U r, a basis of a coarse subspace Build M 1 r I n + U r ( λ n T 1 I r )U T r, T = UT r BUr Apply (m) to B M 1 r At each restart, update r and the basis U r D A Adaptive D(m,r) Detect a potential slow convergence [Sosonkina et al 1998] Switch to D(m,r) only if necessary [Nuentsa Wakam et al 2013] 8 / 21
19 Implementation of Deflated in PETSc New KSP type : D D A Usage in Petsc Available in PETSc Use D just as any other KSP with the following options 9 / 21
20 Experiments with CFD matrices FLUOREM matrices in MatrixMarket collection large, sparse, nonsymmetric matrices linearization of Navier-Stokes: symmetric profile with structured blocks Schwarz preconditioning combined with or D [Nuentsa Wakam+Erhel+Gropp 2013; Nuentsa Wakam+Pacull 2013] D A RM07R n= 381,689; nnz=37,464,962 HV15R n= 2,017,169; nnz=283,073, / 21
21 Combining D with domain decomposition RM07R HV15R D A CPU time on parallel computers RM07R, n = 381,689, nz = 37,464,962 D FULL- (48) D(32,2,15) ITS Time (s) ITS Time (s) ITS Time (s) r HV15R, n = 2,017,169, nz = 283,073,458 D FULL- (150) D(96,10,70) ITS Time (s) ITS Time (s) ITS Time (s) r (8.4E-08) 3, (6.8E-01) / 21
22 A: deflation with an adaptive augmented basis Building blocks Initial step: run one cycle of (m) and compute shifts for the Newton basis Compute U r = [u 0, u 1,..., u r 1 ] a basis of a coarse subspace First step: build a basis K m+1 = [k 0, k 1,..., k m] of the Krylov subspace K m+1 (B, r 0 ) such that D A BK m = K m+1 T m Define the augmented subspace C s = K m(b, r 0 ) + span{u r } with s = m + r with the basis [ Km U r ] compute BU r = ˆK r D r Define the augmented subspace Ĉ s+1 = K m+1 (B, r 0 ) + span{bu r } with the basis [ Km+1 ˆK r ] 12 / 21
23 A: deflation with an adaptive augmented basis Building blocks Initial step: run one cycle of (m) and compute shifts for the Newton basis Compute U r = [u 0, u 1,..., u r 1 ] a basis of a coarse subspace First step: build a basis K m+1 = [k 0, k 1,..., k m] of the Krylov subspace K m+1 (B, r 0 ) such that D A BK m = K m+1 T m Define the augmented subspace C s = K m(b, r 0 ) + span{u r } with s = m + r with the basis [ Km U r ] compute BU r = ˆK r D r Define the augmented subspace Ĉ s+1 = K m+1 (B, r 0 ) + span{bu r } with the basis [ Km+1 ˆK r ] 12 / 21
24 A: deflation with an adaptive augmented basis Building blocks Second step: Compute an orthonormal basis of Ĉ s+1 QR factorize the augmented basis [ K m+1 ˆK r ] = Vs+1 R s+1 BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m D A BU r = (V m+1 R m+1,r + V r R r )D r Define the basis W s = [ V m U r ] BW s = V s+1 H s Third step: x s = x 0 + M 1 W s y s r s = r 0 BW s y s = V s+1 (βe 1 H s y s ) and β = r 0 2 y s = min y R s βe 1 H s y 2 Final step: Adaptively update r and the coarse basis U r 13 / 21
25 A: deflation with an adaptive augmented basis Building blocks Second step: Compute an orthonormal basis of Ĉ s+1 QR factorize the augmented basis [ K m+1 ˆK r ] = Vs+1 R s+1 BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m D A BU r = (V m+1 R m+1,r + V r R r )D r Define the basis W s = [ V m U r ] BW s = V s+1 H s Third step: x s = x 0 + M 1 W s y s r s = r 0 BW s y s = V s+1 (βe 1 H s y s ) and β = r 0 2 y s = min y R s βe 1 H s y 2 Final step: Adaptively update r and the coarse basis U r 13 / 21
26 A: deflation with an adaptive augmented basis Building blocks Second step: Compute an orthonormal basis of Ĉ s+1 QR factorize the augmented basis [ K m+1 ˆK r ] = Vs+1 R s+1 BK m = V m+1 R m+1 T m BV m = V m+1 R m+1 T mr 1 m D A BU r = (V m+1 R m+1,r + V r R r )D r Define the basis W s = [ V m U r ] BW s = V s+1 H s Third step: x s = x 0 + M 1 W s y s r s = r 0 BW s y s = V s+1 (βe 1 H s y s ) and β = r 0 2 y s = min y R s βe 1 H s y 2 Final step: Adaptively update r and the coarse basis U r 13 / 21
27 A in PETSc New KSP type : A figure D A Usage in Petsc Use A just as KSPSetType(ksp, KSPA) or -ksp type agmres, -pc type asm,... Options : -ksp gmres restart m, -ksp agmres eig r, -ksp max its maxits, -ksp agmres smv smv -ksp agmres bgv bgv, / 21
28 Experiments with Augmented RM07R: effect of the restarting RM07R: effect of the number of subdomains RM07R-A(m,2)-(m) RM07R-(m)-A(m,2)-D (32) A(32,2) (48) A(48,2) (32)-D32 A(32,2)-D32 (32)-D64 A(32,2)-D64 D A Rel. Res. Norm e-06 Rel. Res. Norm e-06 1e-08 1e-08 1e e Number of Iterations Number of Iterations CPU time on parallel computers RM07R, n = 381,689, nz = 37,464,962 D (32) A(32,r) ITS Time (s) ITS Time (s) / 21
29 Parallel CPU Time with A Convection-Diffusion test cases 3DCONSKY 121 : size = 1,771,561; nonzeros = 50,178,241 3DCOSKY 161 : size= 4,173,281; nonzeros = 118,645,121 [Nuentsa Wakam+Erhel 2014] D A 3DCONSKY 121 3DCONSKY 161 CPU Time DCONSKY_121 (16) A(16,1) Subdomains CPU Time DCONSKY_161 (16) A(16,1) Subdomains 16 / 21
30 Augmented Conjugate Gradient A Symmetric Positive Definite (SPD) matrix is a Krylov method short recurrences and minimization properties preconditioning M 1 SIDNUR Coarse grid and balancing [Nicolaides 1987, Mandel 1993, DD proceedings, Giraud et al.] Z basis of a coarse subspace restriction of A: nonsingular small matrix A c = Z T AZ projections: P = I AZA 1 c Z T and P T = I ZA 1 c (AZ) T coarse grid: ZA 1 c Z T balancing: C b = P T M 1 P + ZA 1 c Z T Coarse grid and augmented CG [Erhel et al 2000, Saad et al. 2000, Tang et al. 2009, Poirriez 2011, Nataf et al] x 0 = ZA 1 c Z T b C a = P T M 1 C a is equivalent to C b (if no loss of orthogonality) 17 / 21
31 SIDNUR: AugCG and Domain Decomposition Balancing Neumann Neumann applied to a Schur complement Neumann-Neumann preconditioning M 1 Balancing with a coarse grid Z SIDNUR SIDNUR [Poirriez 2011, Pichot et al. 2014] domain decomposition provided by the user coarse grid : signature of subdomains [Frank and Vuik 2001] C++ library mutual factorization of local Schur complements and local matrices management of floating subdomains numerical experiments with 3D fracture networks 18 / 21
32 Experiments with SIDNUR Flow computations in Discrete Fracture Networks random domain generated with MPFRAC software SPD sparse matrix solving with SIDNUR [Poirriez 2011] SIDNUR Number of iterations CPU time S c h u r S c h u r N b o f ite ra tio n s C P U T im e x x x x x x x x x x s y s te m s iz e n 6 x x x x x x x x x x s y s te m s iz e 19 / 21
33 Comparing SIDNUR with other solvers UMFPACK: direct solver Boomer-AMG: algebraic multigrid : Conjugate Gradient preconditioned by Boomer-AMG [Poirriez 2011] SIDNUR CPU time P C G A M G U m f S c h u r re la tiv e C P U tim e 1 0,1 6 x x x x x x x x x x s y s te m s iz e 20 / 21
34 Conclusion D KSP module: deflation in (m) with or without Newton basis A KSP module: augmented Newton basis in (m) Deflation combined with Schwarz domain decomposition preconditioning Robustness: reduce the restarting effects and the domain decomposition effects Efficiency: increase granularity and scalability Numerical experiments with CFD problems: D and A faster than SIDNUR Deflation combined with Schur domain decomposition SIDNUR: Balancing Domain Decomposition Robustness: reduce the domain decomposition effects Efficiency: parallel Schur and Neumann Neumann computations Numerical experiments with 3D fracture networks: faster than multigrid and Library soon available as free software 21 / 21
C. Vuik 1 R. Nabben 2 J.M. Tang 1
Deflation acceleration of block ILU preconditioned Krylov methods C. Vuik 1 R. Nabben 2 J.M. Tang 1 1 Delft University of Technology J.M. Burgerscentrum 2 Technical University Berlin Institut für Mathematik
More informationOUTLINE ffl CFD: elliptic pde's! Ax = b ffl Basic iterative methods ffl Krylov subspace methods ffl Preconditioning techniques: Iterative methods ILU
Preconditioning Techniques for Solving Large Sparse Linear Systems Arnold Reusken Institut für Geometrie und Praktische Mathematik RWTH-Aachen OUTLINE ffl CFD: elliptic pde's! Ax = b ffl Basic iterative
More informationOn the choice of abstract projection vectors for second level preconditioners
On the choice of abstract projection vectors for second level preconditioners C. Vuik 1, J.M. Tang 1, and R. Nabben 2 1 Delft University of Technology 2 Technische Universität Berlin Institut für Mathematik
More informationContents. Preface... xi. Introduction...
Contents Preface... xi Introduction... xv Chapter 1. Computer Architectures... 1 1.1. Different types of parallelism... 1 1.1.1. Overlap, concurrency and parallelism... 1 1.1.2. Temporal and spatial parallelism
More informationLecture 8: Fast Linear Solvers (Part 7)
Lecture 8: Fast Linear Solvers (Part 7) 1 Modified Gram-Schmidt Process with Reorthogonalization Test Reorthogonalization If Av k 2 + δ v k+1 2 = Av k 2 to working precision. δ = 10 3 2 Householder Arnoldi
More informationWHEN studying distributed simulations of power systems,
1096 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL 21, NO 3, AUGUST 2006 A Jacobian-Free Newton-GMRES(m) Method with Adaptive Preconditioner and Its Application for Power Flow Calculations Ying Chen and Chen
More informationPreface to the Second Edition. Preface to the First Edition
n page v Preface to the Second Edition Preface to the First Edition xiii xvii 1 Background in Linear Algebra 1 1.1 Matrices................................. 1 1.2 Square Matrices and Eigenvalues....................
More informationarxiv: v2 [math.na] 1 Feb 2013
A FRAMEWORK FOR DEFLATED AND AUGMENTED KRYLOV SUBSPACE METHODS ANDRÉ GAUL, MARTIN H. GUTKNECHT, JÖRG LIESEN AND REINHARD NABBEN arxiv:1206.1506v2 [math.na] 1 Feb 2013 Abstract. We consider deflation and
More informationArnoldi Methods in SLEPc
Scalable Library for Eigenvalue Problem Computations SLEPc Technical Report STR-4 Available at http://slepc.upv.es Arnoldi Methods in SLEPc V. Hernández J. E. Román A. Tomás V. Vidal Last update: October,
More informationLinear Solvers. Andrew Hazel
Linear Solvers Andrew Hazel Introduction Thus far we have talked about the formulation and discretisation of physical problems...... and stopped when we got to a discrete linear system of equations. Introduction
More informationMultilevel spectral coarse space methods in FreeFem++ on parallel architectures
Multilevel spectral coarse space methods in FreeFem++ on parallel architectures Pierre Jolivet Laboratoire Jacques-Louis Lions Laboratoire Jean Kuntzmann DD 21, Rennes. June 29, 2012 In collaboration with
More informationOptimal Left and Right Additive Schwarz Preconditioning for Minimal Residual Methods with Euclidean and Energy Norms
Optimal Left and Right Additive Schwarz Preconditioning for Minimal Residual Methods with Euclidean and Energy Norms Marcus Sarkis Worcester Polytechnic Inst., Mass. and IMPA, Rio de Janeiro and Daniel
More informationFrom Direct to Iterative Substructuring: some Parallel Experiences in 2 and 3D
From Direct to Iterative Substructuring: some Parallel Experiences in 2 and 3D Luc Giraud N7-IRIT, Toulouse MUMPS Day October 24, 2006, ENS-INRIA, Lyon, France Outline 1 General Framework 2 The direct
More informationA High-Performance Parallel Hybrid Method for Large Sparse Linear Systems
Outline A High-Performance Parallel Hybrid Method for Large Sparse Linear Systems Azzam Haidar CERFACS, Toulouse joint work with Luc Giraud (N7-IRIT, France) and Layne Watson (Virginia Polytechnic Institute,
More informationSparse linear solvers: iterative methods and preconditioning
Sparse linear solvers: iterative methods and preconditioning L Grigori ALPINES INRIA and LJLL, UPMC March 2017 Plan Sparse linear solvers Sparse matrices and graphs Classes of linear solvers Krylov subspace
More informationSolving Large Nonlinear Sparse Systems
Solving Large Nonlinear Sparse Systems Fred W. Wubs and Jonas Thies Computational Mechanics & Numerical Mathematics University of Groningen, the Netherlands f.w.wubs@rug.nl Centre for Interdisciplinary
More informationAMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 23: GMRES and Other Krylov Subspace Methods; Preconditioning
AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 23: GMRES and Other Krylov Subspace Methods; Preconditioning Xiangmin Jiao SUNY Stony Brook Xiangmin Jiao Numerical Analysis I 1 / 18 Outline
More informationFast Iterative Solution of Saddle Point Problems
Michele Benzi Department of Mathematics and Computer Science Emory University Atlanta, GA Acknowledgments NSF (Computational Mathematics) Maxim Olshanskii (Mech-Math, Moscow State U.) Zhen Wang (PhD student,
More information4.8 Arnoldi Iteration, Krylov Subspaces and GMRES
48 Arnoldi Iteration, Krylov Subspaces and GMRES We start with the problem of using a similarity transformation to convert an n n matrix A to upper Hessenberg form H, ie, A = QHQ, (30) with an appropriate
More informationAdaptive preconditioners for nonlinear systems of equations
Adaptive preconditioners for nonlinear systems of equations L. Loghin D. Ruiz A. Touhami CERFACS Technical Report TR/PA/04/42 Also ENSEEIHT-IRIT Technical Report RT/TLSE/04/02 Abstract The use of preconditioned
More informationInexactness and flexibility in linear Krylov solvers
Inexactness and flexibility in linear Krylov solvers Luc Giraud ENSEEIHT (N7) - IRIT, Toulouse Matrix Analysis and Applications CIRM Luminy - October 15-19, 2007 in honor of Gérard Meurant for his 60 th
More informationPreconditioners for the incompressible Navier Stokes equations
Preconditioners for the incompressible Navier Stokes equations C. Vuik M. ur Rehman A. Segal Delft Institute of Applied Mathematics, TU Delft, The Netherlands SIAM Conference on Computational Science and
More informationDELFT UNIVERSITY OF TECHNOLOGY
DELFT UNIVERSITY OF TECHNOLOGY REPORT 16-02 The Induced Dimension Reduction method applied to convection-diffusion-reaction problems R. Astudillo and M. B. van Gijzen ISSN 1389-6520 Reports of the Delft
More informationLecture 17: Iterative Methods and Sparse Linear Algebra
Lecture 17: Iterative Methods and Sparse Linear Algebra David Bindel 25 Mar 2014 Logistics HW 3 extended to Wednesday after break HW 4 should come out Monday after break Still need project description
More informationThe Deflation Accelerated Schwarz Method for CFD
The Deflation Accelerated Schwarz Method for CFD J. Verkaik 1, C. Vuik 2,, B.D. Paarhuis 1, and A. Twerda 1 1 TNO Science and Industry, Stieltjesweg 1, P.O. Box 155, 2600 AD Delft, The Netherlands 2 Delft
More informationSummary of Iterative Methods for Non-symmetric Linear Equations That Are Related to the Conjugate Gradient (CG) Method
Summary of Iterative Methods for Non-symmetric Linear Equations That Are Related to the Conjugate Gradient (CG) Method Leslie Foster 11-5-2012 We will discuss the FOM (full orthogonalization method), CG,
More informationSOLVING SPARSE LINEAR SYSTEMS OF EQUATIONS. Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA
1 SOLVING SPARSE LINEAR SYSTEMS OF EQUATIONS Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA 2 OUTLINE Sparse matrix storage format Basic factorization
More informationParallel Numerics, WT 2016/ Iterative Methods for Sparse Linear Systems of Equations. page 1 of 1
Parallel Numerics, WT 2016/2017 5 Iterative Methods for Sparse Linear Systems of Equations page 1 of 1 Contents 1 Introduction 1.1 Computer Science Aspects 1.2 Numerical Problems 1.3 Graphs 1.4 Loop Manipulations
More informationFEM and sparse linear system solving
FEM & sparse linear system solving, Lecture 9, Nov 19, 2017 1/36 Lecture 9, Nov 17, 2017: Krylov space methods http://people.inf.ethz.ch/arbenz/fem17 Peter Arbenz Computer Science Department, ETH Zürich
More informationA Domain Decomposition Based Jacobi-Davidson Algorithm for Quantum Dot Simulation
A Domain Decomposition Based Jacobi-Davidson Algorithm for Quantum Dot Simulation Tao Zhao 1, Feng-Nan Hwang 2 and Xiao-Chuan Cai 3 Abstract In this paper, we develop an overlapping domain decomposition
More informationPreliminary Results of GRAPES Helmholtz solver using GCR and PETSc tools
Preliminary Results of GRAPES Helmholtz solver using GCR and PETSc tools Xiangjun Wu (1),Lilun Zhang (2),Junqiang Song (2) and Dehui Chen (1) (1) Center for Numerical Weather Prediction, CMA (2) School
More informationPreconditioning Techniques for Large Linear Systems Part III: General-Purpose Algebraic Preconditioners
Preconditioning Techniques for Large Linear Systems Part III: General-Purpose Algebraic Preconditioners Michele Benzi Department of Mathematics and Computer Science Emory University Atlanta, Georgia, USA
More informationOn deflation and singular symmetric positive semi-definite matrices
Journal of Computational and Applied Mathematics 206 (2007) 603 614 www.elsevier.com/locate/cam On deflation and singular symmetric positive semi-definite matrices J.M. Tang, C. Vuik Faculty of Electrical
More informationEIGIFP: A MATLAB Program for Solving Large Symmetric Generalized Eigenvalue Problems
EIGIFP: A MATLAB Program for Solving Large Symmetric Generalized Eigenvalue Problems JAMES H. MONEY and QIANG YE UNIVERSITY OF KENTUCKY eigifp is a MATLAB program for computing a few extreme eigenvalues
More informationITERATIVE METHODS FOR SPARSE LINEAR SYSTEMS
ITERATIVE METHODS FOR SPARSE LINEAR SYSTEMS YOUSEF SAAD University of Minnesota PWS PUBLISHING COMPANY I(T)P An International Thomson Publishing Company BOSTON ALBANY BONN CINCINNATI DETROIT LONDON MADRID
More informationCommunication-avoiding Krylov subspace methods
Motivation Communication-avoiding Krylov subspace methods Mark mhoemmen@cs.berkeley.edu University of California Berkeley EECS MS Numerical Libraries Group visit: 28 April 2008 Overview Motivation Current
More informationOn solving linear systems arising from Shishkin mesh discretizations
On solving linear systems arising from Shishkin mesh discretizations Petr Tichý Faculty of Mathematics and Physics, Charles University joint work with Carlos Echeverría, Jörg Liesen, and Daniel Szyld October
More informationThe amount of work to construct each new guess from the previous one should be a small multiple of the number of nonzeros in A.
AMSC/CMSC 661 Scientific Computing II Spring 2005 Solution of Sparse Linear Systems Part 2: Iterative methods Dianne P. O Leary c 2005 Solving Sparse Linear Systems: Iterative methods The plan: Iterative
More informationRobust solution of Poisson-like problems with aggregation-based AMG
Robust solution of Poisson-like problems with aggregation-based AMG Yvan Notay Université Libre de Bruxelles Service de Métrologie Nucléaire Paris, January 26, 215 Supported by the Belgian FNRS http://homepages.ulb.ac.be/
More informationA Robust Preconditioned Iterative Method for the Navier-Stokes Equations with High Reynolds Numbers
Applied and Computational Mathematics 2017; 6(4): 202-207 http://www.sciencepublishinggroup.com/j/acm doi: 10.11648/j.acm.20170604.18 ISSN: 2328-5605 (Print); ISSN: 2328-5613 (Online) A Robust Preconditioned
More informationParallelism in FreeFem++.
Parallelism in FreeFem++. Guy Atenekeng 1 Frederic Hecht 2 Laura Grigori 1 Jacques Morice 2 Frederic Nataf 2 1 INRIA, Saclay 2 University of Paris 6 Workshop on FreeFem++, 2009 Outline 1 Introduction Motivation
More informationMaster Thesis Literature Study Presentation
Master Thesis Literature Study Presentation Delft University of Technology The Faculty of Electrical Engineering, Mathematics and Computer Science January 29, 2010 Plaxis Introduction Plaxis Finite Element
More informationAMS Mathematics Subject Classification : 65F10,65F50. Key words and phrases: ILUS factorization, preconditioning, Schur complement, 1.
J. Appl. Math. & Computing Vol. 15(2004), No. 1, pp. 299-312 BILUS: A BLOCK VERSION OF ILUS FACTORIZATION DAVOD KHOJASTEH SALKUYEH AND FAEZEH TOUTOUNIAN Abstract. ILUS factorization has many desirable
More informationPreconditioned inverse iteration and shift-invert Arnoldi method
Preconditioned inverse iteration and shift-invert Arnoldi method Melina Freitag Department of Mathematical Sciences University of Bath CSC Seminar Max-Planck-Institute for Dynamics of Complex Technical
More informationIndefinite and physics-based preconditioning
Indefinite and physics-based preconditioning Jed Brown VAW, ETH Zürich 2009-01-29 Newton iteration Standard form of a nonlinear system F (u) 0 Iteration Solve: Update: J(ũ)u F (ũ) ũ + ũ + u Example (p-bratu)
More informationAn advanced ILU preconditioner for the incompressible Navier-Stokes equations
An advanced ILU preconditioner for the incompressible Navier-Stokes equations M. ur Rehman C. Vuik A. Segal Delft Institute of Applied Mathematics, TU delft The Netherlands Computational Methods with Applications,
More informationMARCH 24-27, 2014 SAN JOSE, CA
MARCH 24-27, 2014 SAN JOSE, CA Sparse HPC on modern architectures Important scientific applications rely on sparse linear algebra HPCG a new benchmark proposal to complement Top500 (HPL) To solve A x =
More informationToward black-box adaptive domain decomposition methods
Toward black-box adaptive domain decomposition methods Frédéric Nataf Laboratory J.L. Lions (LJLL), CNRS, Alpines Inria and Univ. Paris VI joint work with Victorita Dolean (Univ. Nice Sophia-Antipolis)
More informationEfficient Deflation for Communication-Avoiding Krylov Subspace Methods
Efficient Deflation for Communication-Avoiding Krylov Subspace Methods Erin Carson Nicholas Knight, James Demmel Univ. of California, Berkeley Monday, June 24, NASCA 2013, Calais, France Overview We derive
More informationMultipréconditionnement adaptatif pour les méthodes de décomposition de domaine. Nicole Spillane (CNRS, CMAP, École Polytechnique)
Multipréconditionnement adaptatif pour les méthodes de décomposition de domaine Nicole Spillane (CNRS, CMAP, École Polytechnique) C. Bovet (ONERA), P. Gosselet (ENS Cachan), A. Parret Fréaud (SafranTech),
More informationAMS526: Numerical Analysis I (Numerical Linear Algebra)
AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 23: GMRES and Other Krylov Subspace Methods Xiangmin Jiao SUNY Stony Brook Xiangmin Jiao Numerical Analysis I 1 / 9 Minimizing Residual CG
More informationDeflation Strategies to Improve the Convergence of Communication-Avoiding GMRES
Deflation Strategies to Improve the Convergence of Communication-Avoiding GMRES Ichitaro Yamazaki, Stanimire Tomov, and Jack Dongarra Department of Electrical Engineering and Computer Science University
More informationM.A. Botchev. September 5, 2014
Rome-Moscow school of Matrix Methods and Applied Linear Algebra 2014 A short introduction to Krylov subspaces for linear systems, matrix functions and inexact Newton methods. Plan and exercises. M.A. Botchev
More informationIntegration of PETSc for Nonlinear Solves
Integration of PETSc for Nonlinear Solves Ben Jamroz, Travis Austin, Srinath Vadlamani, Scott Kruger Tech-X Corporation jamroz@txcorp.com http://www.txcorp.com NIMROD Meeting: Aug 10, 2010 Boulder, CO
More informationPROJECTED GMRES AND ITS VARIANTS
PROJECTED GMRES AND ITS VARIANTS Reinaldo Astudillo Brígida Molina rastudillo@kuaimare.ciens.ucv.ve bmolina@kuaimare.ciens.ucv.ve Centro de Cálculo Científico y Tecnológico (CCCT), Facultad de Ciencias,
More informationNumerical behavior of inexact linear solvers
Numerical behavior of inexact linear solvers Miro Rozložník joint results with Zhong-zhi Bai and Pavel Jiránek Institute of Computer Science, Czech Academy of Sciences, Prague, Czech Republic The fourth
More informationScalable Domain Decomposition Preconditioners For Heterogeneous Elliptic Problems
Scalable Domain Decomposition Preconditioners For Heterogeneous Elliptic Problems Pierre Jolivet, F. Hecht, F. Nataf, C. Prud homme Laboratoire Jacques-Louis Lions Laboratoire Jean Kuntzmann INRIA Rocquencourt
More informationLast Time. Social Network Graphs Betweenness. Graph Laplacian. Girvan-Newman Algorithm. Spectral Bisection
Eigenvalue Problems Last Time Social Network Graphs Betweenness Girvan-Newman Algorithm Graph Laplacian Spectral Bisection λ 2, w 2 Today Small deviation into eigenvalue problems Formulation Standard eigenvalue
More informationSolving Sparse Linear Systems: Iterative methods
Scientific Computing with Case Studies SIAM Press, 2009 http://www.cs.umd.edu/users/oleary/sccs Lecture Notes for Unit VII Sparse Matrix Computations Part 2: Iterative Methods Dianne P. O Leary c 2008,2010
More informationSolving Sparse Linear Systems: Iterative methods
Scientific Computing with Case Studies SIAM Press, 2009 http://www.cs.umd.edu/users/oleary/sccswebpage Lecture Notes for Unit VII Sparse Matrix Computations Part 2: Iterative Methods Dianne P. O Leary
More informationDeflation and augmentation techniques in Krylov subspace methods for the solution of linear systems
arxiv:1303.5692v1 [math.na] 21 Mar 2013 Deflation and augmentation techniques in Krylov subspace methods for the solution of linear systems Olivier Coulaud, Luc Giraud, Pierre Ramet, Xavier Vasseur RESEARCH
More informationPreconditioned GMRES Revisited
Preconditioned GMRES Revisited Roland Herzog Kirk Soodhalter UBC (visiting) RICAM Linz Preconditioning Conference 2017 Vancouver August 01, 2017 Preconditioned GMRES Revisited Vancouver 1 / 32 Table of
More informationThe Lanczos and conjugate gradient algorithms
The Lanczos and conjugate gradient algorithms Gérard MEURANT October, 2008 1 The Lanczos algorithm 2 The Lanczos algorithm in finite precision 3 The nonsymmetric Lanczos algorithm 4 The Golub Kahan bidiagonalization
More informationTWO-GRID DEFLATED KRYLOV METHODS FOR LINEAR EQUATIONS
1 TWO-GRID DEFLATED KRYLOV METHODS FOR LINEAR EQUATIONS RONALD B. MORGAN AND ZHAO YANG Abstract. An approach is given for solving large linear systems that combines Krylov methods with use of two different
More informationAPPLIED NUMERICAL LINEAR ALGEBRA
APPLIED NUMERICAL LINEAR ALGEBRA James W. Demmel University of California Berkeley, California Society for Industrial and Applied Mathematics Philadelphia Contents Preface 1 Introduction 1 1.1 Basic Notation
More informationThe solution of the discretized incompressible Navier-Stokes equations with iterative methods
The solution of the discretized incompressible Navier-Stokes equations with iterative methods Report 93-54 C. Vuik Technische Universiteit Delft Delft University of Technology Faculteit der Technische
More information9.1 Preconditioned Krylov Subspace Methods
Chapter 9 PRECONDITIONING 9.1 Preconditioned Krylov Subspace Methods 9.2 Preconditioned Conjugate Gradient 9.3 Preconditioned Generalized Minimal Residual 9.4 Relaxation Method Preconditioners 9.5 Incomplete
More informationAlgebraic Multigrid as Solvers and as Preconditioner
Ò Algebraic Multigrid as Solvers and as Preconditioner Domenico Lahaye domenico.lahaye@cs.kuleuven.ac.be http://www.cs.kuleuven.ac.be/ domenico/ Department of Computer Science Katholieke Universiteit Leuven
More informationOn the Superlinear Convergence of MINRES. Valeria Simoncini and Daniel B. Szyld. Report January 2012
On the Superlinear Convergence of MINRES Valeria Simoncini and Daniel B. Szyld Report 12-01-11 January 2012 This report is available in the World Wide Web at http://www.math.temple.edu/~szyld 0 Chapter
More informationUne méthode parallèle hybride à deux niveaux interfacée dans un logiciel d éléments finis
Une méthode parallèle hybride à deux niveaux interfacée dans un logiciel d éléments finis Frédéric Nataf Laboratory J.L. Lions (LJLL), CNRS, Alpines Inria and Univ. Paris VI Victorita Dolean (Univ. Nice
More informationUsing an Auction Algorithm in AMG based on Maximum Weighted Matching in Matrix Graphs
Using an Auction Algorithm in AMG based on Maximum Weighted Matching in Matrix Graphs Pasqua D Ambra Institute for Applied Computing (IAC) National Research Council of Italy (CNR) pasqua.dambra@cnr.it
More informationDepartment of Computer Science, University of Illinois at Urbana-Champaign
Department of Computer Science, University of Illinois at Urbana-Champaign Probing for Schur Complements and Preconditioning Generalized Saddle-Point Problems Eric de Sturler, sturler@cs.uiuc.edu, http://www-faculty.cs.uiuc.edu/~sturler
More informationFast solvers for steady incompressible flow
ICFD 25 p.1/21 Fast solvers for steady incompressible flow Andy Wathen Oxford University wathen@comlab.ox.ac.uk http://web.comlab.ox.ac.uk/~wathen/ Joint work with: Howard Elman (University of Maryland,
More informationTopics. The CG Algorithm Algorithmic Options CG s Two Main Convergence Theorems
Topics The CG Algorithm Algorithmic Options CG s Two Main Convergence Theorems What about non-spd systems? Methods requiring small history Methods requiring large history Summary of solvers 1 / 52 Conjugate
More informationNumerical Methods in Matrix Computations
Ake Bjorck Numerical Methods in Matrix Computations Springer Contents 1 Direct Methods for Linear Systems 1 1.1 Elements of Matrix Theory 1 1.1.1 Matrix Algebra 2 1.1.2 Vector Spaces 6 1.1.3 Submatrices
More informationIterative Methods for Sparse Linear Systems
Iterative Methods for Sparse Linear Systems Luca Bergamaschi e-mail: berga@dmsa.unipd.it - http://www.dmsa.unipd.it/ berga Department of Mathematical Methods and Models for Scientific Applications University
More informationLecture 3: Inexact inverse iteration with preconditioning
Lecture 3: Department of Mathematical Sciences CLAPDE, Durham, July 2008 Joint work with M. Freitag (Bath), and M. Robbé & M. Sadkane (Brest) 1 Introduction 2 Preconditioned GMRES for Inverse Power Method
More informationSolving Ax = b, an overview. Program
Numerical Linear Algebra Improving iterative solvers: preconditioning, deflation, numerical software and parallelisation Gerard Sleijpen and Martin van Gijzen November 29, 27 Solving Ax = b, an overview
More informationAMS526: Numerical Analysis I (Numerical Linear Algebra)
AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 19: More on Arnoldi Iteration; Lanczos Iteration Xiangmin Jiao Stony Brook University Xiangmin Jiao Numerical Analysis I 1 / 17 Outline 1
More informationConvergence Behavior of a Two-Level Optimized Schwarz Preconditioner
Convergence Behavior of a Two-Level Optimized Schwarz Preconditioner Olivier Dubois 1 and Martin J. Gander 2 1 IMA, University of Minnesota, 207 Church St. SE, Minneapolis, MN 55455 dubois@ima.umn.edu
More informationCONVERGENCE BOUNDS FOR PRECONDITIONED GMRES USING ELEMENT-BY-ELEMENT ESTIMATES OF THE FIELD OF VALUES
European Conference on Computational Fluid Dynamics ECCOMAS CFD 2006 P. Wesseling, E. Oñate and J. Périaux (Eds) c TU Delft, The Netherlands, 2006 CONVERGENCE BOUNDS FOR PRECONDITIONED GMRES USING ELEMENT-BY-ELEMENT
More informationConjugate gradient method. Descent method. Conjugate search direction. Conjugate Gradient Algorithm (294)
Conjugate gradient method Descent method Hestenes, Stiefel 1952 For A N N SPD In exact arithmetic, solves in N steps In real arithmetic No guaranteed stopping Often converges in many fewer than N steps
More informationNumerical Methods - Numerical Linear Algebra
Numerical Methods - Numerical Linear Algebra Y. K. Goh Universiti Tunku Abdul Rahman 2013 Y. K. Goh (UTAR) Numerical Methods - Numerical Linear Algebra I 2013 1 / 62 Outline 1 Motivation 2 Solving Linear
More informationA hybrid reordered Arnoldi method to accelerate PageRank computations
A hybrid reordered Arnoldi method to accelerate PageRank computations Danielle Parker Final Presentation Background Modeling the Web The Web The Graph (A) Ranks of Web pages v = v 1... Dominant Eigenvector
More informationANR Project DEDALES Algebraic and Geometric Domain Decomposition for Subsurface Flow
ANR Project DEDALES Algebraic and Geometric Domain Decomposition for Subsurface Flow Michel Kern Inria Paris Rocquencourt Maison de la Simulation C2S@Exa Days, Inria Paris Centre, Novembre 2016 M. Kern
More informationConjugate Gradients: Idea
Overview Steepest Descent often takes steps in the same direction as earlier steps Wouldn t it be better every time we take a step to get it exactly right the first time? Again, in general we choose a
More informationRecent advances in approximation using Krylov subspaces. V. Simoncini. Dipartimento di Matematica, Università di Bologna.
Recent advances in approximation using Krylov subspaces V. Simoncini Dipartimento di Matematica, Università di Bologna and CIRSA, Ravenna, Italy valeria@dm.unibo.it 1 The framework It is given an operator
More informationarxiv: v4 [math.na] 1 Sep 2018
A Comparison of Preconditioned Krylov Subspace Methods for Large-Scale Nonsymmetric Linear Systems Aditi Ghai, Cao Lu and Xiangmin Jiao arxiv:1607.00351v4 [math.na] 1 Sep 2018 September 5, 2018 Department
More informationKey words. linear equations, eigenvalues, polynomial preconditioning, GMRES, GMRES-DR, deflation, QCD
1 POLYNOMIAL PRECONDITIONED GMRES AND GMRES-DR QUAN LIU, RONALD B. MORGAN, AND WALTER WILCOX Abstract. We look at solving large nonsymmetric systems of linear equations using polynomial preconditioned
More informationPreconditioning Techniques Analysis for CG Method
Preconditioning Techniques Analysis for CG Method Huaguang Song Department of Computer Science University of California, Davis hso@ucdavis.edu Abstract Matrix computation issue for solve linear system
More information1 Conjugate gradients
Notes for 2016-11-18 1 Conjugate gradients We now turn to the method of conjugate gradients (CG), perhaps the best known of the Krylov subspace solvers. The CG iteration can be characterized as the iteration
More informationA decade of fast and robust Helmholtz solvers
A decade of fast and robust Helmholtz solvers Werkgemeenschap Scientific Computing Spring meeting Kees Vuik May 11th, 212 1 Delft University of Technology Contents Introduction Preconditioning (22-28)
More informationOn the Preconditioning of the Block Tridiagonal Linear System of Equations
On the Preconditioning of the Block Tridiagonal Linear System of Equations Davod Khojasteh Salkuyeh Department of Mathematics, University of Mohaghegh Ardabili, PO Box 179, Ardabil, Iran E-mail: khojaste@umaacir
More informationDELFT UNIVERSITY OF TECHNOLOGY
DELFT UNIVERSITY OF TECHNOLOGY REPORT 06-05 Solution of the incompressible Navier Stokes equations with preconditioned Krylov subspace methods M. ur Rehman, C. Vuik G. Segal ISSN 1389-6520 Reports of the
More informationApplied Mathematics 205. Unit V: Eigenvalue Problems. Lecturer: Dr. David Knezevic
Applied Mathematics 205 Unit V: Eigenvalue Problems Lecturer: Dr. David Knezevic Unit V: Eigenvalue Problems Chapter V.4: Krylov Subspace Methods 2 / 51 Krylov Subspace Methods In this chapter we give
More informationCourse Notes: Week 1
Course Notes: Week 1 Math 270C: Applied Numerical Linear Algebra 1 Lecture 1: Introduction (3/28/11) We will focus on iterative methods for solving linear systems of equations (and some discussion of eigenvalues
More informationDeflated Krylov Iterations in Domain Decomposition Methods
305 Deflated Krylov Iterations in Domain Decomposition Methods Y.L.Gurieva 1, V.P.Ilin 1,2, and D.V.Perevozkin 1 1 Introduction The goal of this research is an investigation of some advanced versions of
More informationAccelerated Solvers for CFD
Accelerated Solvers for CFD Co-Design of Hardware/Software for Predicting MAV Aerodynamics Eric de Sturler, Virginia Tech Mathematics Email: sturler@vt.edu Web: http://www.math.vt.edu/people/sturler Co-design
More informationInexact inverse iteration with preconditioning
Department of Mathematical Sciences Computational Methods with Applications Harrachov, Czech Republic 24th August 2007 (joint work with M. Robbé and M. Sadkane (Brest)) 1 Introduction 2 Preconditioned
More informationA communication-avoiding thick-restart Lanczos method on a distributed-memory system
A communication-avoiding thick-restart Lanczos method on a distributed-memory system Ichitaro Yamazaki and Kesheng Wu Lawrence Berkeley National Laboratory, Berkeley, CA, USA Abstract. The Thick-Restart
More information