Challenges for Matrix Preconditioning Methods

Size: px
Start display at page:

Download "Challenges for Matrix Preconditioning Methods"

Transcription

1 Challenges for Matrix Preconditioning Methods Matthias Bollhoefer 1 1 Dept. of Mathematics TU Berlin Preconditioning 2005, Atlanta, May 19, 2005 supported by the DFG research center MATHEON in Berlin

2 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

3 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

4 Circuit and device simulation Circuits Modified nodal approach transient analysis Devices harmonic balance drift diffusion equations

5 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

6 The Anderson model of localization A Challenge for modern eigenvalue algorithms Model describes electronic transport properties in disordered systems Wave function probabilities ω = 14.5 ω = 16.5 ω = 18.5

7 The Anderson model of localization A Challenge for modern eigenvalue algorithms Uniform grid in a three dimensional cube 1 1 ε ijk x ijk P p i =1, q j =1, r j =1 xpqr = λx ijk ε ijk [ ω/2, ω/2], ω radomly in [1, 30]. Eigenvalue problem Ax = λx in 3D x quantum mechanic wave function ω 16.5 critical range ω 16.5 fluctuations, but bounded ω 16.5 wave functions are localized

8 Challenges Physically sensible results require large scale simulation, n = m 3 (e.g. m = 100, 200,... ) physically interesting: eigenvectors around λ = 0 at ω c = 16.5 requires eigenvalue solver which compute some eigenvectors around λ = 0 at ω c are fast (some hours up to a few days) are memory efficient (required memory scales n)

9 Challenges Physically sensible results require large scale simulation, n = m 3 (e.g. m = 100, 200,... ) physically interesting: eigenvectors around λ = 0 at ω c = 16.5 requires eigenvalue solver which compute some eigenvectors around λ = 0 at ω c are fast (some hours up to a few days) are memory efficient (required memory scales n) Numerical Approach Preconditioned Eigenvalue Solver

10 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

11 What are Matchings Example 0 original A 1 C A

12 What are Matchings Example 0 original A 1 C A Rows Columns Associated graph

13 What are Matchings Example 0 original A 1 C A Rows Columns Associated graph (Perfect) matching. (At most) one edge for each row/column

14 What are Matchings Example 0 original A 1 C A Rows Columns (Perfect) matching. (At most) one edge for each row/column

15 What are Matchings Example 0 original A 1 C A Rows Columns π = (5, 3, 4, 2, 1) Associated permutation

16 What are Matchings Example 0 original A 1 C A Rows Columns π = (5, 3, 4, 2, 1) 0 permuted Π A 1 C A Permuted matrix

17 Minimum Weight Matchings Matching Permutation of the matrix to zero free representation [ MC21, Duff 77] Refined Objective: Strengthening diagonal dominance ny Find matching such that a π(i),i is maximized i=1

18 Minimum Weight Matchings Matching Permutation of the matrix to zero free representation [ MC21, Duff 77] Refined Objective: Strengthening diagonal dominance ny Find matching such that a π(i),i is maximized i=1 dense case [Olschowka, Neumeier 96], sparse case [ MC64, Duff,Koster 99]

19 Minimum Weight Matchings nx reformulation: minimize c π(i),i, where c ij = j i=1 maxj a ij log a ij a ij 0 otherwise known linear sum assignment problem

20 Minimum Weight Matchings nx reformulation: minimize c π(i),i, where c ij = j i=1 maxj a ij log a ij a ij 0 otherwise known linear sum assignment problem solution leads to two vectors (u i ) i, (v j ) j such that u i + u j c ij and if (i, j) are part of the matching u i + u j = c ij diagonal scalings D r = dgl (e u i ) i, D c = dgl (e v i / max k a jk ) j

21 Minimum Weight Matchings nx reformulation: minimize c π(i),i, where c ij = j i=1 maxj a ij log a ij a ij 0 otherwise known linear sum assignment problem solution leads to two vectors (u i ) i, (v j ) j such that u i + u j c ij and if (i, j) are part of the matching u i + u j = c ij diagonal scalings D r = dgl (e u i ) i, D c = dgl (e v i / max k a jk ) j yields permutation Π and diagonal scalings D r and D c. A Π D r AD c such that a ii = 1, a ij 1 for all i, j

22 How matchings improve iterative solvers Example 33 large sparse unstructured systems from chemical engineering Simple dual threshold ILU with pivoting from SPARSKIT [Saad 94] Restarted GMRES(30), check convergence after at most 500 steps AMF MMD RCM ND ILUTP, percentage of problems solved without matching 70 percentage e 3 1e 4 1e 5 drop tolerance

23 How matchings improve iterative solvers Example 33 large sparse unstructured systems from chemical engineering Simple dual threshold ILU with pivoting from SPARSKIT [Saad 94] Restarted GMRES(30), check convergence after at most 500 steps ILUTP, percentage of problems solved WITH MC64 AMF MMD RCM ND with matching 70 percentage e 3 1e 4 1e 5 drop tolerance

24 How matchings improve iterative solvers Minimum weight matchings significantly improve preconditioning methods [Benzi,Haws,Tuma 00] also applicable to factored sparse approximate inverses High potential also as part direct solvers (PARDISO, SuperLU,... ) Matchings allow Level 3 BLAS with static pivoting use of symmetric reordering techniques

25 How matchings improve iterative solvers BUT Minimum weight matchings significantly improve preconditioning methods [Benzi,Haws,Tuma 00] also applicable to factored sparse approximate inverses High potential also as part direct solvers (PARDISO, SuperLU,... ) Matchings allow Level 3 BLAS with static pivoting use of symmetric reordering techniques matching destroys symmetry structures leads to unsymmetric scaling in this way not directly applicable to symmetrically structured systems

26 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

27 Basic Idea of Symmetric Matchings Symmetric matchings [Duff,Gilbert 02], [Duff,Pralet 04] Example 0 1 C A π = (5, 3, 4, 2, 1) original A

28 Basic Idea of Symmetric Matchings Symmetric matchings [Duff,Gilbert 02], [Duff,Pralet 04] Decompose permutation as product of cycles Example 0 1 C A π = (5, 3, 4, 2, 1) = (5, 1)(3, 4, 2) original A

29 Basic Idea of Symmetric Matchings Symmetric matchings [Duff,Gilbert 02], [Duff,Pralet 04] Decompose permutation as product of cycles each cycle is associated with a diagonal block, pivots stay there Example 0 1 C A π = (5, 3, 4, 2, 1) = (5, 1)(3, 4, 2) 0 1 C A original A permuted Π A

30 Basic Idea of Symmetric Matchings Symmetric matchings [Duff,Gilbert 02], [Duff,Pralet 04] Decompose permutation as product of cycles each cycle is associated with a diagonal block, pivots stay there pivots stay inside if related SYMMETRIC permutation is applied Example 0 1 C A π = (5, 3, 4, 2, 1) = (5, 1)(3, 4, 2) 0 1 C A original A permuted Π AΠ

31 Basic Idea of Symmetric Matchings break up cycles into pairs of 2 cycles and group them together various strategies of breaking up the cycles [Duff,Gilbert 02],[Duff,Pralet 04],[Schenk,Hagemann 04] long cycles rarely show up in practice [Duff,Pralet 04] Scaling replaced by symmetric scaling D = (D cd r ) 1/2 ensures that entries are still at most 1 in modulus Example 0 1 C A 0 1 C A 0 1 C A Π AΠ rearranged Π 1 AΠ 1 final Π 2 AΠ 2

32 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

33 Symbolic reorderings techniques multilevel nested dissection MeTiS [Karypis,Kumar 95] Approximate minimum degree [Amestoy,Davis,Duff,Gilbert,Larimore,Ng] AMD, UMFPACK further old fashioned orderings (Reverse Cuthill McKee, Minimum Degree) Different approach: diagonal dominance + sparsity [Saad 03] alternative to matching + reordering

34 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

35 Algebraic Multilevel Approach j F fine grid points Reordering (+ rescaling) the system C coarse grid points «A Π AFF A FC AΠ = A CF A CC

36 Algebraic Multilevel Approach j F fine grid points Reordering (+ rescaling) the system C coarse grid points «A Π AFF A FC AΠ = Approximate block decomposition ««««AFF A FC LFF 0 DFF 0 UFF U FC = + E A CF A CC L CF I 0 S CC 0 I {z } L A CF A CC {z } D A {z } U

37 Algebraic Multilevel Approach j F fine grid points Reordering (+ rescaling) the system C coarse grid points «A Π AFF A FC AΠ = Approximate block decomposition ««««AFF A FC LFF 0 DFF 0 UFF U FC = + E A CF A CC L CF I 0 S CC 0 I {z } L A CF A CC {z } D A {z } U S CC coarse grid system, E error matrix E represents the entries being discarded in L, U

38 solution operator Approximation B FF A 1 FF, B FC A 1 FF AFC, B CF A CFA 1 FF, e.g. via solving with LFFDFFUFF e.g. via L 1 FF LFC e.g. via UCFU 1 FF

39 solution operator Approximation B FF A 1 FF, B FC A 1 FF AFC, B CF A CFA 1 FF, e.g. via solving with LFFDFFUFF e.g. via L 1 FF LFC e.g. via UCFU 1 FF AFF A FC A CF A CC «1 BFF «BFC + I «{z } P S 1 ` CC BCF {z I } R P interpolation R restriction

40 solution operator Approximation B FF A 1 FF, B FC A 1 FF AFC, B CF A CFA 1 FF, e.g. via solving with LFFDFFUFF e.g. via L 1 FF LFC e.g. via UCFU 1 FF AFF A FC A CF A CC «1 BFF «BFC + I «{z } P S 1 ` CC BCF {z I } R P interpolation R restriction Multilevel approach: same approach recursively applied to S CC

41 Solution Operator Refinement Supplement by smoothing steps G 1, G 2 (e.g. Jacobi, Gauss Seidel) Iterations matrix for the error e = x x j «ff BFF 0 e (I + PS 1 CC R A)e 0 0 e (I G 2 A)(I PS 1 CC R A)(I G 1 A)e

42 Solution Operator Refinement Supplement by smoothing steps G 1, G 2 (e.g. Jacobi, Gauss Seidel) Iterations matrix for the error e = x x j «ff BFF 0 e (I + PS 1 CC R A)e 0 0 e (I G 2 A)(I PS 1 CC R A)(I G 1 A)e V cycle (µ = 1), W cycle (µ = 2) (I G 2 A)(I PS 1 CC R A) µ (I G 1 A)

43 What means inverse based decomposition Prescribed uniform bound κ for the inverse transformations L 1, U 1 L 1 = L 1 FF 0 L CFL 1 FF «I κ U 1 = U 1 FF U 1 FF UFC 0 I «κ

44 What means inverse based decomposition Prescribed uniform bound κ for the inverse transformations L 1, U 1 L 1 = L 1 FF 0 L CFL 1 FF I «L 1 FF 0 B CF I R «κ U 1 = U 1 FF U 1 FF UFC 0 I «B FC U 1 FF 0 I P! κ

45 Why inverse based Decompositions Generally speaking: Norm of the inverse factors drive the approximation error Inverse error F = L 1 EU 1 is amplified If the norms of inverse factors are even kept bounded: Absolute error of the coarse grid system S CC can be predicted Tight κ forces approximately sparse factors L 1, U 1

46 Approximation Error Why Inverse Based Decompositions Approximation A = LDU + E For solving Ax = b we have to apply L 1, U 1

47 Approximation Error Why Inverse Based Decompositions Approximation L 1 AU 1 = D + F Can be used directly to construct an approximate inverse decomposition

48 Approximation Error Why Inverse Based Decompositions Approximation L 1 AU 1 = D + F Can be used directly to construct an approximate inverse decomposition AINV [Benzi, Tuma 98] compute approximate inverses W L 1, Z U 1 directly without L, U A!! + F 1 + F 2 WAZ =! + F Numerical usually more expensive than ILU Significantly more robust if small entries in W, Z are discarded. ILUs as by-product from AINV inherit robustness [Benzi,Tuma 03]

49 Approximation Error Why Inverse Based Decompositions w ij z kl ff ε j wij := 0 z kl := 0 Theorem [B.,Saad 02] l im e m W u mj Ze m ) ( lim := 0 ε u mj := 0 e i I LW e m (m i)ε, e m I ZU e j (m j)ε

50 Approximation Error Why Inverse Based Decompositions w ij z kl ff ε j wij := 0 z kl := 0 Theorem [B.,Saad 02] l im e m W u mj Ze m ) ( lim := 0 ε u mj := 0 e i I LW e m (m i)ε, e m I ZU e j (m j)ε Observation Norm of the inverse factors drive the approximation error

51 The Inverse Error Why Inverse Based Decompositions LFF 0 I L CF «1 UFF U FC A 0 I «1 DFF 0 = 0 S CC «! F + FC F CF {z } F

52 The Inverse Error Why Inverse Based Decompositions LFF 0 I L CF «1 UFF U FC A 0 I «1 DFF 0 = 0 S CC «! F + FC F CF {z } F Lemma [B.,Saad 04] Denote by E L,FF, E U,FF the entries being dropped from L FF, U FF. coarse grid system S CC from ILU F FC = L 1 FF (E L,FF D FF + D FFE U,FF ) U 1 FFU FC Coarse grid system S CC = R AP (successively obtained via Galerkin) F FC = D FFE U,FF U 1 FFU FC

53 Error Propagation Why Inverse Based Decompositoins Suppose that the diagonal entries occuring during the decomposition L FFD FFU FF = A FF + E FF are uniformly bounded.

54 Error Propagation Why Inverse Based Decompositoins Suppose that the diagonal entries occuring during the decomposition L FFD FFU FF = A FF + E FF are uniformly bounded. Theorem [B.,Saad 04] Coarse grid system S CC from ILU If l im, u mj are dropped, whenever l im, u mj ε/κ 2, There exists a constant K such that s ij s ij K (κε) 2 Coarse grid system S CC = R AP (successively from Galerkin) If l im, u mj are dropped, whenever l im, u mj ε, There exists a constant K such that s ij s ij K (κε) 2

55 Consequences for inverse based multilevel methods Estimate L 1, U 1 efficiently [Cline,Moler,Stewart,Wilkinson 77] Construct a well suited initial system by a priori permutation and scaling ( A FF) Keep L 1, U 1 below κ by inverse-based pivoting factor postpone Tight κ desired, since L 1 κ P j<i (L 1 ) ij κ 1

56 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

57 Aggressive Dropping Problem for practical problems we do not precisely know the optimal ε to be save we prefer a smaller tolerance Consequences as ε 0, the fill in significantly increases as ε 0, number of iteration steps decreases to a few number of steps memory requirement dramatically increases even for the iterative part, efficiency does not increase since the fill increases

58 Aggressive Dropping Solution Lemma We do not necessarily need a small number of iteration steps instead: L 1 AU 1 = D F should lead to small F and small perturbations in L 1, U 1 could be tolerated Denote by µ k, ν k the number of nonzeros in column k of L (resp. row k of U). Let L, Ũ be constructed from L, U by dropping entries l ik, u kj satisfying L 1 e i l ik τ µ k, u kj e j U 1 τ ν k, then where D F = (I + E L )( D F )(I + E U ) E L 1 τ, E U τ.

59 Aggressive Dropping Example. BCSSTK25. Compare time and memory WITHOUT and WITH aggressive dropping drop tolerance ε is decreased threshold τ for aggressive dropping is kept at time[sec] 10 9 nnz(l)/nnz(a) drop tolerance drop tolerance 10 6

60 Aggressive Dropping Example. BCSSTK25. Compare time and memory WITHOUT and WITH aggressive dropping drop tolerance ε is decreased threshold τ for aggressive dropping is kept at time[sec] 10 9 nnz(l)/nnz(a) drop tolerance drop tolerance 10 6

61 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

62 Circuit Simulation (benchmark collection from Infineon) compare regular and inverse based ILU Dependence of the decomposition on drop tolerance ε (Convergence of GMRES(30) after at most 500 steps) simple ILU inverse based ILU 6 6 number of systems number of systems e 3 1e e 3 1e 4 drop tolerance ε drop tolerance ε

63 Device Simulation direct solver (PARDISO, [Schenk,Gärner 04] excellent, but causes a lot of fill inverse based ILU (ILUPACK [B.,Saad 04]): fix drop tolerance at ε = 10 3 and use κ = 10. regular ILU fails until 1e 7 for most problems both use minimum weight matching and MeTiS. direct solver inverse based ILU nnz(l+u) nnz(l+u) time[sec] nnz(a) nnz(a) time[sec] barrier e e1 barrier e e1 barrier e3 0.9 barrier e e2 barrier e e1 barrier e e1 barrier e e1 barrier e e1

64 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

65 Challenges Symmetry requires structure preserving solver Block oriented variant (1 1, 2 2), since highly indefinite Matching based preprocessing [Duff,Koster 98], [Gilbert,Duff 02], [Duff,Pralet 04], [Schenk,Hagemann 04] classical approach: Cullum Willoughby algorithm [ 85] iterative methods ARPACK [Lehoucq,Sorensen,Yang 98], Jacobi Davidson [Sleijpen,Van der Vorst 96],[Geus 02] direct solver PARDISO [Schenk,Gärtner], preconditioner ILUPACK [B.,Saad 04],[B.,Schenk 05]

66 Numerical results [Schenk,B.,Römer 05] n = m 3, m = 100, 130, 160, 190. ω = ω c = (SGI Altix/BX2, Intel Itanium2, 1.6GHz) Time system size n = m Cullum Willoughby 71 : 04

67 Numerical results [Schenk,B.,Römer 05] n = m 3, m = 100, 130, 160, 190. ω = ω c = (SGI Altix/BX2, Intel Itanium2, 1.6GHz) Time system size n = m Cullum Willoughby 71 : 04 ARPACK + PARDISO 5 : 37

68 Numerical results [Schenk,B.,Römer 05] n = m 3, m = 100, 130, 160, 190. ω = ω c = (SGI Altix/BX2, Intel Itanium2, 1.6GHz) Time system size n = m Cullum Willoughby 71 : 04 ARPACK + PARDISO 5 : 37 ARPACK + ILUPACK 0 : 45 5 : : 58

69 Numerical results [Schenk,B.,Römer 05] n = m 3, m = 100, 130, 160, 190. ω = ω c = (SGI Altix/BX2, Intel Itanium2, 1.6GHz) Time system size n = m Cullum Willoughby 71 : 04 ARPACK + PARDISO 5 : 37 ARPACK + ILUPACK 0 : 45 5 : : 58 Jacobi Davidson + ILUPACK 0 : 18 1 : 01 5 : 03 9 : 06

70 Numerical results [Schenk,B.,Römer, 05] n = m 3, m = 100, 130, 160, 190. ω = ω c = Memory [GB] system size n = m ARPACK + PARDISO 14.3 ARPACK + ILUPACK Jacobi Davidson + ILUPACK

71 Inverse based decompositions: the key to success clear: modern iterative eigenvalue solver (Jacobi-Davidson) essential Matching quite helpful key role: bound κ for L 1 Numerical example m = 70, 100. Vary bound κ inside ILUPACK fill in nnz(l) nnz(a) κ(m = 70) κ(m = 100) Time [sec] 1.7e2 3.7e2 7.8e2 5.2e2 1.1e3

72 Outline 1 Challenging application problems Circuit and device simulation Anderson Model a case study 2 Advances in Preconditioning Techniques Matchings Symmetric Matchings Symbolic reorderings techniques Inverse-based Techniques Aggressive Dropping 3 Application Problems Circuit and Device Simulation Anderson Model a case study ILUPACK

73 ILUPACK V2.0 multilevel ILU preconditioning software package inverse based decompositions single/double, real and complex arithmetic supported classes of matrices that are supported general symmetric (Hermitian) positive definite new in V2.0 real/complex symmetric, complex Hermitian indefinite interfaces to incorporate matchings is provided (PARDISO, MC64) to be released soon...

74 Conclusions development in recent years has significantly changed preconditioning methods as well as direct solvers matchings dramatically stabilize preconditioning methods symmetric matchings nowadays allow to efficiently treat symmetrically structured indefinite systems inverse based decompositions are complementary approach analysis partially gives an explanation for this effect

J.I. Aliaga 1 M. Bollhöfer 2 A.F. Martín 1 E.S. Quintana-Ortí 1. March, 2009

J.I. Aliaga 1 M. Bollhöfer 2 A.F. Martín 1 E.S. Quintana-Ortí 1. March, 2009 Parallel Preconditioning of Linear Systems based on ILUPACK for Multithreaded Architectures J.I. Aliaga M. Bollhöfer 2 A.F. Martín E.S. Quintana-Ortí Deparment of Computer Science and Engineering, Univ.

More information

Scientific Computing

Scientific Computing Scientific Computing Direct solution methods Martin van Gijzen Delft University of Technology October 3, 2018 1 Program October 3 Matrix norms LU decomposition Basic algorithm Cost Stability Pivoting Pivoting

More information

Preconditioning Techniques for Large Linear Systems Part III: General-Purpose Algebraic Preconditioners

Preconditioning 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 information

Parallelization of Multilevel Preconditioners Constructed from Inverse-Based ILUs on Shared-Memory Multiprocessors

Parallelization of Multilevel Preconditioners Constructed from Inverse-Based ILUs on Shared-Memory Multiprocessors Parallelization of Multilevel Preconditioners Constructed from Inverse-Based ILUs on Shared-Memory Multiprocessors J.I. Aliaga 1 M. Bollhöfer 2 A.F. Martín 1 E.S. Quintana-Ortí 1 1 Deparment of Computer

More information

Adaptive algebraic multigrid methods in lattice computations

Adaptive algebraic multigrid methods in lattice computations Adaptive algebraic multigrid methods in lattice computations Karsten Kahl Bergische Universität Wuppertal January 8, 2009 Acknowledgements Matthias Bolten, University of Wuppertal Achi Brandt, Weizmann

More information

AMS526: Numerical Analysis I (Numerical Linear Algebra for Computational and Data Sciences)

AMS526: Numerical Analysis I (Numerical Linear Algebra for Computational and Data Sciences) AMS526: Numerical Analysis I (Numerical Linear Algebra for Computational and Data Sciences) Lecture 19: Computing the SVD; Sparse Linear Systems Xiangmin Jiao Stony Brook University Xiangmin Jiao Numerical

More information

Preconditioners for the incompressible Navier Stokes equations

Preconditioners 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 information

Recent advances in sparse linear solver technology for semiconductor device simulation matrices

Recent advances in sparse linear solver technology for semiconductor device simulation matrices Recent advances in sparse linear solver technology for semiconductor device simulation matrices (Invited Paper) Olaf Schenk and Michael Hagemann Department of Computer Science University of Basel Basel,

More information

Scientific Computing WS 2018/2019. Lecture 9. Jürgen Fuhrmann Lecture 9 Slide 1

Scientific Computing WS 2018/2019. Lecture 9. Jürgen Fuhrmann Lecture 9 Slide 1 Scientific Computing WS 2018/2019 Lecture 9 Jürgen Fuhrmann juergen.fuhrmann@wias-berlin.de Lecture 9 Slide 1 Lecture 9 Slide 2 Simple iteration with preconditioning Idea: Aû = b iterative scheme û = û

More information

Numerical Linear Algebra

Numerical Linear Algebra Numerical Linear Algebra Decompositions, numerical aspects Gerard Sleijpen and Martin van Gijzen September 27, 2017 1 Delft University of Technology Program Lecture 2 LU-decomposition Basic algorithm Cost

More information

Program Lecture 2. Numerical Linear Algebra. Gaussian elimination (2) Gaussian elimination. Decompositions, numerical aspects

Program Lecture 2. Numerical Linear Algebra. Gaussian elimination (2) Gaussian elimination. Decompositions, numerical aspects Numerical Linear Algebra Decompositions, numerical aspects Program Lecture 2 LU-decomposition Basic algorithm Cost Stability Pivoting Cholesky decomposition Sparse matrices and reorderings Gerard Sleijpen

More information

An advanced ILU preconditioner for the incompressible Navier-Stokes equations

An 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 information

Scientific Computing with Case Studies SIAM Press, Lecture Notes for Unit VII Sparse Matrix

Scientific Computing with Case Studies SIAM Press, Lecture Notes for Unit VII Sparse Matrix 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 1: Direct Methods Dianne P. O Leary c 2008

More information

SOLVING SPARSE LINEAR SYSTEMS OF EQUATIONS. Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA

SOLVING 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 information

Maximum-weighted matching strategies and the application to symmetric indefinite systems

Maximum-weighted matching strategies and the application to symmetric indefinite systems Maximum-weighted matching strategies and the application to symmetric indefinite systems by Stefan Röllin, and Olaf Schenk 2 Technical Report CS-24-7 Department of Computer Science, University of Basel

More information

Incomplete Cholesky preconditioners that exploit the low-rank property

Incomplete Cholesky preconditioners that exploit the low-rank property anapov@ulb.ac.be ; http://homepages.ulb.ac.be/ anapov/ 1 / 35 Incomplete Cholesky preconditioners that exploit the low-rank property (theory and practice) Artem Napov Service de Métrologie Nucléaire, Université

More information

Direct solution methods for sparse matrices. p. 1/49

Direct solution methods for sparse matrices. p. 1/49 Direct solution methods for sparse matrices p. 1/49 p. 2/49 Direct solution methods for sparse matrices Solve Ax = b, where A(n n). (1) Factorize A = LU, L lower-triangular, U upper-triangular. (2) Solve

More information

Preface to the Second Edition. Preface to the First Edition

Preface 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 information

Multigrid Methods and their application in CFD

Multigrid Methods and their application in CFD Multigrid Methods and their application in CFD Michael Wurst TU München 16.06.2009 1 Multigrid Methods Definition Multigrid (MG) methods in numerical analysis are a group of algorithms for solving differential

More information

OUTLINE ffl CFD: elliptic pde's! Ax = b ffl Basic iterative methods ffl Krylov subspace methods ffl Preconditioning techniques: Iterative methods ILU

OUTLINE 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 information

Numerical Methods I Non-Square and Sparse Linear Systems

Numerical Methods I Non-Square and Sparse Linear Systems Numerical Methods I Non-Square and Sparse Linear Systems Aleksandar Donev Courant Institute, NYU 1 donev@courant.nyu.edu 1 MATH-GA 2011.003 / CSCI-GA 2945.003, Fall 2014 September 25th, 2014 A. Donev (Courant

More information

Incomplete LU Preconditioning and Error Compensation Strategies for Sparse Matrices

Incomplete LU Preconditioning and Error Compensation Strategies for Sparse Matrices Incomplete LU Preconditioning and Error Compensation Strategies for Sparse Matrices Eun-Joo Lee Department of Computer Science, East Stroudsburg University of Pennsylvania, 327 Science and Technology Center,

More information

Robust solution of Poisson-like problems with aggregation-based AMG

Robust 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 information

A robust multilevel approximate inverse preconditioner for symmetric positive definite matrices

A robust multilevel approximate inverse preconditioner for symmetric positive definite matrices DICEA DEPARTMENT OF CIVIL, ENVIRONMENTAL AND ARCHITECTURAL ENGINEERING PhD SCHOOL CIVIL AND ENVIRONMENTAL ENGINEERING SCIENCES XXX CYCLE A robust multilevel approximate inverse preconditioner for symmetric

More information

Stabilization and Acceleration of Algebraic Multigrid Method

Stabilization and Acceleration of Algebraic Multigrid Method Stabilization and Acceleration of Algebraic Multigrid Method Recursive Projection Algorithm A. Jemcov J.P. Maruszewski Fluent Inc. October 24, 2006 Outline 1 Need for Algorithm Stabilization and Acceleration

More information

6. Iterative Methods for Linear Systems. The stepwise approach to the solution...

6. Iterative Methods for Linear Systems. The stepwise approach to the solution... 6 Iterative Methods for Linear Systems The stepwise approach to the solution Miriam Mehl: 6 Iterative Methods for Linear Systems The stepwise approach to the solution, January 18, 2013 1 61 Large Sparse

More information

Fine-grained Parallel Incomplete LU Factorization

Fine-grained Parallel Incomplete LU Factorization Fine-grained Parallel Incomplete LU Factorization Edmond Chow School of Computational Science and Engineering Georgia Institute of Technology Sparse Days Meeting at CERFACS June 5-6, 2014 Contribution

More information

A greedy strategy for coarse-grid selection

A greedy strategy for coarse-grid selection A greedy strategy for coarse-grid selection S. MacLachlan Yousef Saad August 3, 2006 Abstract Efficient solution of the very large linear systems that arise in numerical modelling of real-world applications

More information

5.1 Banded Storage. u = temperature. The five-point difference operator. uh (x, y + h) 2u h (x, y)+u h (x, y h) uh (x + h, y) 2u h (x, y)+u h (x h, y)

5.1 Banded Storage. u = temperature. The five-point difference operator. uh (x, y + h) 2u h (x, y)+u h (x, y h) uh (x + h, y) 2u h (x, y)+u h (x h, y) 5.1 Banded Storage u = temperature u= u h temperature at gridpoints u h = 1 u= Laplace s equation u= h u = u h = grid size u=1 The five-point difference operator 1 u h =1 uh (x + h, y) 2u h (x, y)+u h

More information

Chapter 7 Iterative Techniques in Matrix Algebra

Chapter 7 Iterative Techniques in Matrix Algebra Chapter 7 Iterative Techniques in Matrix Algebra Per-Olof Persson persson@berkeley.edu Department of Mathematics University of California, Berkeley Math 128B Numerical Analysis Vector Norms Definition

More information

Preconditioning techniques to accelerate the convergence of the iterative solution methods

Preconditioning techniques to accelerate the convergence of the iterative solution methods Note Preconditioning techniques to accelerate the convergence of the iterative solution methods Many issues related to iterative solution of linear systems of equations are contradictory: numerical efficiency

More information

Using 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 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 information

FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION

FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION EDMOND CHOW AND AFTAB PATEL Abstract. This paper presents a new fine-grained parallel algorithm for computing an incomplete LU factorization. All nonzeros

More information

Fine-Grained Parallel Algorithms for Incomplete Factorization Preconditioning

Fine-Grained Parallel Algorithms for Incomplete Factorization Preconditioning Fine-Grained Parallel Algorithms for Incomplete Factorization Preconditioning Edmond Chow School of Computational Science and Engineering Georgia Institute of Technology, USA SPPEXA Symposium TU München,

More information

Solving linear systems (6 lectures)

Solving linear systems (6 lectures) Chapter 2 Solving linear systems (6 lectures) 2.1 Solving linear systems: LU factorization (1 lectures) Reference: [Trefethen, Bau III] Lecture 20, 21 How do you solve Ax = b? (2.1.1) In numerical linear

More information

Incomplete factorization preconditioners and their updates with applications - I 1,2

Incomplete factorization preconditioners and their updates with applications - I 1,2 Incomplete factorization preconditioners and their updates with applications - I 1,2 Daniele Bertaccini, Fabio Durastante Moscow August 24, 216 Notes of the course: Incomplete factorization preconditioners

More information

Aggregation-based algebraic multigrid

Aggregation-based algebraic multigrid Aggregation-based algebraic multigrid from theory to fast solvers Yvan Notay Université Libre de Bruxelles Service de Métrologie Nucléaire CEMRACS, Marseille, July 18, 2012 Supported by the Belgian FNRS

More information

Solving Large Nonlinear Sparse Systems

Solving 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 information

JADAMILU: a software code for computing selected eigenvalues of large sparse symmetric matrices

JADAMILU: a software code for computing selected eigenvalues of large sparse symmetric matrices JADAMILU: a software code for computing selected eigenvalues of large sparse symmetric matrices Matthias Bollhöfer 1 Institut Computational Mathematics, Technische Universität Braunschweig, D 38106 Braunschweig,

More information

Solving Ax = b, an overview. Program

Solving 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 information

1. Fast Iterative Solvers of SLE

1. Fast Iterative Solvers of SLE 1. Fast Iterative Solvers of crucial drawback of solvers discussed so far: they become slower if we discretize more accurate! now: look for possible remedies relaxation: explicit application of the multigrid

More information

Lecture 8: Fast Linear Solvers (Part 7)

Lecture 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 information

c 2000 Society for Industrial and Applied Mathematics

c 2000 Society for Industrial and Applied Mathematics SIAM J. SCI. COMPUT. Vol. 22, No. 4, pp. 1333 1353 c 2000 Society for Industrial and Applied Mathematics PRECONDITIONING HIGHLY INDEFINITE AND NONSYMMETRIC MATRICES MICHELE BENZI, JOHN C. HAWS, AND MIROSLAV

More information

Solving Symmetric Indefinite Systems with Symmetric Positive Definite Preconditioners

Solving Symmetric Indefinite Systems with Symmetric Positive Definite Preconditioners Solving Symmetric Indefinite Systems with Symmetric Positive Definite Preconditioners Eugene Vecharynski 1 Andrew Knyazev 2 1 Department of Computer Science and Engineering University of Minnesota 2 Department

More information

Jordan Journal of Mathematics and Statistics (JJMS) 5(3), 2012, pp A NEW ITERATIVE METHOD FOR SOLVING LINEAR SYSTEMS OF EQUATIONS

Jordan Journal of Mathematics and Statistics (JJMS) 5(3), 2012, pp A NEW ITERATIVE METHOD FOR SOLVING LINEAR SYSTEMS OF EQUATIONS Jordan Journal of Mathematics and Statistics JJMS) 53), 2012, pp.169-184 A NEW ITERATIVE METHOD FOR SOLVING LINEAR SYSTEMS OF EQUATIONS ADEL H. AL-RABTAH Abstract. The Jacobi and Gauss-Seidel iterative

More information

Boundary Value Problems - Solving 3-D Finite-Difference problems Jacob White

Boundary Value Problems - Solving 3-D Finite-Difference problems Jacob White Introduction to Simulation - Lecture 2 Boundary Value Problems - Solving 3-D Finite-Difference problems Jacob White Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy Outline Reminder about

More information

Multigrid absolute value preconditioning

Multigrid absolute value preconditioning Multigrid absolute value preconditioning Eugene Vecharynski 1 Andrew Knyazev 2 (speaker) 1 Department of Computer Science and Engineering University of Minnesota 2 Department of Mathematical and Statistical

More information

An efficient multigrid solver based on aggregation

An efficient multigrid solver based on aggregation An efficient multigrid solver based on aggregation Yvan Notay Université Libre de Bruxelles Service de Métrologie Nucléaire Graz, July 4, 2012 Co-worker: Artem Napov Supported by the Belgian FNRS http://homepages.ulb.ac.be/

More information

A Linear Multigrid Preconditioner for the solution of the Navier-Stokes Equations using a Discontinuous Galerkin Discretization. Laslo Tibor Diosady

A Linear Multigrid Preconditioner for the solution of the Navier-Stokes Equations using a Discontinuous Galerkin Discretization. Laslo Tibor Diosady A Linear Multigrid Preconditioner for the solution of the Navier-Stokes Equations using a Discontinuous Galerkin Discretization by Laslo Tibor Diosady B.A.Sc., University of Toronto (2005) Submitted to

More information

Linear Solvers. Andrew Hazel

Linear 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 information

AMG for a Peta-scale Navier Stokes Code

AMG for a Peta-scale Navier Stokes Code AMG for a Peta-scale Navier Stokes Code James Lottes Argonne National Laboratory October 18, 2007 The Challenge Develop an AMG iterative method to solve Poisson 2 u = f discretized on highly irregular

More information

Parallel Preconditioning Methods for Ill-conditioned Problems

Parallel Preconditioning Methods for Ill-conditioned Problems Parallel Preconditioning Methods for Ill-conditioned Problems Kengo Nakajima Information Technology Center, The University of Tokyo 2014 Conference on Advanced Topics and Auto Tuning in High Performance

More information

Krylov Space Solvers

Krylov Space Solvers Seminar for Applied Mathematics ETH Zurich International Symposium on Frontiers of Computational Science Nagoya, 12/13 Dec. 2005 Sparse Matrices Large sparse linear systems of equations or large sparse

More information

Spectral element agglomerate AMGe

Spectral element agglomerate AMGe Spectral element agglomerate AMGe T. Chartier 1, R. Falgout 2, V. E. Henson 2, J. E. Jones 4, T. A. Manteuffel 3, S. F. McCormick 3, J. W. Ruge 3, and P. S. Vassilevski 2 1 Department of Mathematics, Davidson

More information

The flexible incomplete LU preconditioner for large nonsymmetric linear systems. Takatoshi Nakamura Takashi Nodera

The flexible incomplete LU preconditioner for large nonsymmetric linear systems. Takatoshi Nakamura Takashi Nodera Research Report KSTS/RR-15/006 The flexible incomplete LU preconditioner for large nonsymmetric linear systems by Takatoshi Nakamura Takashi Nodera Takatoshi Nakamura School of Fundamental Science and

More information

A Domain Decomposition Based Jacobi-Davidson Algorithm for Quantum Dot Simulation

A 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 information

Mathematics and Computer Science

Mathematics and Computer Science Technical Report TR-2007-002 Block preconditioning for saddle point systems with indefinite (1,1) block by Michele Benzi, Jia Liu Mathematics and Computer Science EMORY UNIVERSITY International Journal

More information

FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION

FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION EDMOND CHOW AND AFTAB PATEL Abstract. This paper presents a new fine-grained parallel algorithm for computing an incomplete LU factorization. All nonzeros

More information

Solving symmetric indefinite systems using memory efficient incomplete factorization preconditioners

Solving symmetric indefinite systems using memory efficient incomplete factorization preconditioners Preprint RAL-P-2015-002 Solving symmetric indefinite systems using memory efficient incomplete factorization preconditioners J Scott, M Tuma, February 2015 Submitted for publication in Siam Journal of

More information

Incomplete Factorization Preconditioners for Least Squares and Linear and Quadratic Programming

Incomplete Factorization Preconditioners for Least Squares and Linear and Quadratic Programming Incomplete Factorization Preconditioners for Least Squares and Linear and Quadratic Programming by Jelena Sirovljevic B.Sc., The University of British Columbia, 2005 A THESIS SUBMITTED IN PARTIAL FULFILMENT

More information

DEN: Linear algebra numerical view (GEM: Gauss elimination method for reducing a full rank matrix to upper-triangular

DEN: Linear algebra numerical view (GEM: Gauss elimination method for reducing a full rank matrix to upper-triangular form) Given: matrix C = (c i,j ) n,m i,j=1 ODE and num math: Linear algebra (N) [lectures] c phabala 2016 DEN: Linear algebra numerical view (GEM: Gauss elimination method for reducing a full rank matrix

More information

c 2000 Society for Industrial and Applied Mathematics

c 2000 Society for Industrial and Applied Mathematics SIAM J. SCI. COMPUT. Vol. 21, No. 5, pp. 1851 1868 c 2000 Society for Industrial and Applied Mathematics ORDERINGS FOR FACTORIZED SPARSE APPROXIMATE INVERSE PRECONDITIONERS MICHELE BENZI AND MIROSLAV TŮMA

More information

Inexact inverse iteration with preconditioning

Inexact 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 information

NEWTON-GMRES PRECONDITIONING FOR DISCONTINUOUS GALERKIN DISCRETIZATIONS OF THE NAVIER-STOKES EQUATIONS

NEWTON-GMRES PRECONDITIONING FOR DISCONTINUOUS GALERKIN DISCRETIZATIONS OF THE NAVIER-STOKES EQUATIONS NEWTON-GMRES PRECONDITIONING FOR DISCONTINUOUS GALERKIN DISCRETIZATIONS OF THE NAVIER-STOKES EQUATIONS P.-O. PERSSON AND J. PERAIRE Abstract. We study preconditioners for the iterative solution of the

More information

Iterative Methods for Solving A x = b

Iterative Methods for Solving A x = b Iterative Methods for Solving A x = b A good (free) online source for iterative methods for solving A x = b is given in the description of a set of iterative solvers called templates found at netlib: http

More information

An Introduction to Algebraic Multigrid (AMG) Algorithms Derrick Cerwinsky and Craig C. Douglas 1/84

An Introduction to Algebraic Multigrid (AMG) Algorithms Derrick Cerwinsky and Craig C. Douglas 1/84 An Introduction to Algebraic Multigrid (AMG) Algorithms Derrick Cerwinsky and Craig C. Douglas 1/84 Introduction Almost all numerical methods for solving PDEs will at some point be reduced to solving A

More information

9.1 Preconditioned Krylov Subspace Methods

9.1 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 information

Solving Linear Systems

Solving Linear Systems Solving Linear Systems Iterative Solutions Methods Philippe B. Laval KSU Fall 207 Philippe B. Laval (KSU) Linear Systems Fall 207 / 2 Introduction We continue looking how to solve linear systems of the

More information

Sparse symmetric preconditioners for dense linear systems in electromagnetism

Sparse symmetric preconditioners for dense linear systems in electromagnetism Sparse symmetric preconditioners for dense linear systems in electromagnetism B. Carpentieri I.S. Duff L. Giraud M. Magolu monga Made CERFACS Technical Report TR/PA//35 Abstract We consider symmetric preconditioning

More information

Max-plus LU. James Hook. November Manchester Institute for Mathematical Sciences School of Mathematics. The University of Manchester

Max-plus LU. James Hook. November Manchester Institute for Mathematical Sciences School of Mathematics. The University of Manchester Max-plus LU James Hook November 25 MIMS EPrint: 25. Manchester Institute for Mathematical Sciences School of Mathematics The University of Manchester Reports available from: And by contacting: http://www.manchester.ac.uk/mims/eprints

More information

Sparse linear solvers

Sparse linear solvers Sparse linear solvers Laura Grigori ALPINES INRIA and LJLL, UPMC On sabbatical at UC Berkeley March 2015 Plan Sparse linear solvers Sparse matrices and graphs Classes of linear solvers Sparse Cholesky

More information

SPARSE SOLVERS POISSON EQUATION. Margreet Nool. November 9, 2015 FOR THE. CWI, Multiscale Dynamics

SPARSE SOLVERS POISSON EQUATION. Margreet Nool. November 9, 2015 FOR THE. CWI, Multiscale Dynamics SPARSE SOLVERS FOR THE POISSON EQUATION Margreet Nool CWI, Multiscale Dynamics November 9, 2015 OUTLINE OF THIS TALK 1 FISHPACK, LAPACK, PARDISO 2 SYSTEM OVERVIEW OF CARTESIUS 3 POISSON EQUATION 4 SOLVERS

More information

Contents. Preface... xi. Introduction...

Contents. 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 information

Algorithm for Sparse Approximate Inverse Preconditioners in the Conjugate Gradient Method

Algorithm for Sparse Approximate Inverse Preconditioners in the Conjugate Gradient Method Algorithm for Sparse Approximate Inverse Preconditioners in the Conjugate Gradient Method Ilya B. Labutin A.A. Trofimuk Institute of Petroleum Geology and Geophysics SB RAS, 3, acad. Koptyug Ave., Novosibirsk

More information

Iterative methods for Linear System

Iterative methods for Linear System Iterative methods for Linear System JASS 2009 Student: Rishi Patil Advisor: Prof. Thomas Huckle Outline Basics: Matrices and their properties Eigenvalues, Condition Number Iterative Methods Direct and

More information

c 2015 Society for Industrial and Applied Mathematics

c 2015 Society for Industrial and Applied Mathematics SIAM J. SCI. COMPUT. Vol. 37, No. 2, pp. C169 C193 c 2015 Society for Industrial and Applied Mathematics FINE-GRAINED PARALLEL INCOMPLETE LU FACTORIZATION EDMOND CHOW AND AFTAB PATEL Abstract. This paper

More information

Linear Algebra. Brigitte Bidégaray-Fesquet. MSIAM, September Univ. Grenoble Alpes, Laboratoire Jean Kuntzmann, Grenoble.

Linear Algebra. Brigitte Bidégaray-Fesquet. MSIAM, September Univ. Grenoble Alpes, Laboratoire Jean Kuntzmann, Grenoble. Brigitte Bidégaray-Fesquet Univ. Grenoble Alpes, Laboratoire Jean Kuntzmann, Grenoble MSIAM, 23 24 September 215 Overview 1 Elementary operations Gram Schmidt orthonormalization Matrix norm Conditioning

More information

Math 471 (Numerical methods) Chapter 3 (second half). System of equations

Math 471 (Numerical methods) Chapter 3 (second half). System of equations Math 47 (Numerical methods) Chapter 3 (second half). System of equations Overlap 3.5 3.8 of Bradie 3.5 LU factorization w/o pivoting. Motivation: ( ) A I Gaussian Elimination (U L ) where U is upper triangular

More information

Using Random Butterfly Transformations to Avoid Pivoting in Sparse Direct Methods

Using Random Butterfly Transformations to Avoid Pivoting in Sparse Direct Methods Using Random Butterfly Transformations to Avoid Pivoting in Sparse Direct Methods Marc Baboulin 1, Xiaoye S. Li 2 and François-Henry Rouet 2 1 University of Paris-Sud, Inria Saclay, France 2 Lawrence Berkeley

More information

AMS526: Numerical Analysis I (Numerical Linear Algebra)

AMS526: Numerical Analysis I (Numerical Linear Algebra) AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 24: Preconditioning and Multigrid Solver Xiangmin Jiao SUNY Stony Brook Xiangmin Jiao Numerical Analysis I 1 / 5 Preconditioning Motivation:

More information

arxiv: v4 [math.na] 1 Sep 2018

arxiv: 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 information

Solving Sparse Linear Systems: Iterative methods

Solving 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 information

Solving Sparse Linear Systems: Iterative methods

Solving 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 information

The amount of work to construct each new guess from the previous one should be a small multiple of the number of nonzeros in A.

The 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 information

Algebraic Multigrid Preconditioners for Computing Stationary Distributions of Markov Processes

Algebraic Multigrid Preconditioners for Computing Stationary Distributions of Markov Processes Algebraic Multigrid Preconditioners for Computing Stationary Distributions of Markov Processes Elena Virnik, TU BERLIN Algebraic Multigrid Preconditioners for Computing Stationary Distributions of Markov

More information

Lecture 17: Iterative Methods and Sparse Linear Algebra

Lecture 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 information

AMS Mathematics Subject Classification : 65F10,65F50. Key words and phrases: ILUS factorization, preconditioning, Schur complement, 1.

AMS 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 information

Verified Solutions of Sparse Linear Systems by LU factorization

Verified Solutions of Sparse Linear Systems by LU factorization Verified Solutions of Sparse Linear Systems by LU factorization TAKESHI OGITA CREST, Japan Science and Technology Agency JST), and Faculty of Science and Engineering, Waseda University, Robert J. Shillman

More information

A dissection solver with kernel detection for unsymmetric matrices in FreeFem++

A dissection solver with kernel detection for unsymmetric matrices in FreeFem++ . p.1/21 11 Dec. 2014, LJLL, Paris FreeFem++ workshop A dissection solver with kernel detection for unsymmetric matrices in FreeFem++ Atsushi Suzuki Atsushi.Suzuki@ann.jussieu.fr Joint work with François-Xavier

More information

ITERATIVE METHODS FOR SPARSE LINEAR SYSTEMS

ITERATIVE 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 information

A short course on: Preconditioned Krylov subspace methods. Yousef Saad University of Minnesota Dept. of Computer Science and Engineering

A short course on: Preconditioned Krylov subspace methods. Yousef Saad University of Minnesota Dept. of Computer Science and Engineering A short course on: Preconditioned Krylov subspace methods Yousef Saad University of Minnesota Dept. of Computer Science and Engineering Universite du Littoral, Jan 19-30, 2005 Outline Part 1 Introd., discretization

More information

LINEAR ALGEBRA: NUMERICAL METHODS. Version: August 12,

LINEAR ALGEBRA: NUMERICAL METHODS. Version: August 12, LINEAR ALGEBRA: NUMERICAL METHODS. Version: August 12, 2000 74 6 Summary Here we summarize the most important information about theoretical and numerical linear algebra. MORALS OF THE STORY: I. Theoretically

More information

Algebraic Multilevel Preconditioner For the Helmholtz Equation In Heterogeneous Media

Algebraic Multilevel Preconditioner For the Helmholtz Equation In Heterogeneous Media Algebraic Multilevel Preconditioner For the Helmholtz Equation In Heterogeneous Media Matthias Bollhöfer Marcus J. Grote Olaf Schenk Abstract An algebraic multilevel (Ml) preconditioner is presented for

More information

Lecture 3: Inexact inverse iteration with preconditioning

Lecture 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 information

A dissection solver with kernel detection for symmetric finite element matrices on shared memory computers

A dissection solver with kernel detection for symmetric finite element matrices on shared memory computers A dissection solver with kernel detection for symmetric finite element matrices on shared memory computers Atsushi Suzuki, François-Xavier Roux To cite this version: Atsushi Suzuki, François-Xavier Roux.

More information

Efficient domain decomposition methods for the time-harmonic Maxwell equations

Efficient domain decomposition methods for the time-harmonic Maxwell equations Efficient domain decomposition methods for the time-harmonic Maxwell equations Marcella Bonazzoli 1, Victorita Dolean 2, Ivan G. Graham 3, Euan A. Spence 3, Pierre-Henri Tournier 4 1 Inria Saclay (Defi

More information

The Conjugate Gradient Method

The Conjugate Gradient Method The Conjugate Gradient Method Classical Iterations We have a problem, We assume that the matrix comes from a discretization of a PDE. The best and most popular model problem is, The matrix will be as large

More information

A Block Compression Algorithm for Computing Preconditioners

A Block Compression Algorithm for Computing Preconditioners A Block Compression Algorithm for Computing Preconditioners J Cerdán, J Marín, and J Mas Abstract To implement efficiently algorithms for the solution of large systems of linear equations in modern computer

More information

Lecture 18 Classical Iterative Methods

Lecture 18 Classical Iterative Methods Lecture 18 Classical Iterative Methods MIT 18.335J / 6.337J Introduction to Numerical Methods Per-Olof Persson November 14, 2006 1 Iterative Methods for Linear Systems Direct methods for solving Ax = b,

More information

A Jacobi Davidson Method with a Multigrid Solver for the Hermitian Wilson-Dirac Operator

A Jacobi Davidson Method with a Multigrid Solver for the Hermitian Wilson-Dirac Operator A Jacobi Davidson Method with a Multigrid Solver for the Hermitian Wilson-Dirac Operator Artur Strebel Bergische Universität Wuppertal August 3, 2016 Joint Work This project is joint work with: Gunnar

More information